{"meta":{"title":"D-Central — GPU ↔ Local-LLM Fit & Sizing","description":"990 GPU×model fit records: for each of the AI-GPU database's cards and each open local LLM, whether the model fits in VRAM at Q4/Q8/FP16, the headroom, and a recommended quantization. Derived by JOIN; weights-first VRAM heuristic.","generated":"2026-06-19T03:22:05+00:00","version":"1.0","license":"https://creativecommons.org/licenses/by/4.0/","license_name":"CC BY 4.0","source":"https://d-central.tech/gpu-llm-compatibility/","record_count":990,"method":"fits = card VRAM >= model weight footprint at quant; quality = tight (<1.15x) / good (<1.6x) / ample (>=1.6x). Real KV-cache + context needs extra VRAM beyond weights — treat \"tight\" as weights-only and size up for long context.","disclaimer":"Estimates derived from published VRAM figures; verify against your runtime (Ollama/llama.cpp/vLLM) before provisioning. Not financial advice."},"rows":[{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"smollm2-1.7b","model_name":"SmolLM2 1.7B Instruct","params_b":1.6999999999999999555910790149937383830547332763671875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":76.400000000000005684341886080801486968994140625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"llama3.2-1b","model_name":"Llama 3.2 1B Instruct","params_b":1,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":77.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"llama3.2-3b","model_name":"Llama 3.2 3B Instruct","params_b":3,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":73.7000000000000028421709430404007434844970703125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"llama3.1-8b","model_name":"Llama 3.1 8B Instruct","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":63.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"llama3.1-70b","model_name":"Llama 3.1 70B Instruct","params_b":70,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":37,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"llama3.3-70b","model_name":"Llama 3.3 70B Instruct","params_b":70,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":37,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"qwen2.5-7b","model_name":"Qwen2.5 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":65.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"qwen2.5-14b","model_name":"Qwen2.5 14B Instruct","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":51.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"qwen2.5-32b","model_name":"Qwen2.5 32B Instruct","params_b":32,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":15,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"qwen2.5-72b","model_name":"Qwen2.5 72B Instruct","params_b":72,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":36,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"qwen3-8b","model_name":"Qwen3 8B","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":63.60000000000000142108547152020037174224853515625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"qwen3-14b","model_name":"Qwen3 14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":51.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"qwen3-32b","model_name":"Qwen3 32B","params_b":32,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":15,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"qwen3-30b-a3b","model_name":"Qwen3 30B-A3B (MoE)","params_b":30,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":18,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"qwen3-235b-a22b","model_name":"Qwen3 235B-A22B (MoE)","params_b":235,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"gemma3-4b","model_name":"Gemma 3 4B Instruct","params_b":4,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":71.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"gemma3-12b","model_name":"Gemma 3 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":55.60000000000000142108547152020037174224853515625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"gemma3-27b","model_name":"Gemma 3 27B Instruct","params_b":27,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":25.300000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"gemma4-e4b","model_name":"Gemma 4 E4B (Edge)","params_b":4.5,"q4_fits":true,"q4_quality":"ample","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"q4","headroom_at_recommended":70.400000000000005684341886080801486968994140625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"gemma4-31b","model_name":"Gemma 4 31B","params_b":31,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":17,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"mistral-7b-v0.3","model_name":"Mistral 7B v0.3 Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":65.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"mistral-nemo-12b","model_name":"Mistral Nemo 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":55.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"mixtral-8x7b","model_name":"Mixtral 8x7B Instruct v0.1","params_b":46.7000000000000028421709430404007434844970703125,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":32.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"phi3.5-mini-3.8b","model_name":"Phi-3.5 mini 3.8B Instruct","params_b":3.79999999999999982236431605997495353221893310546875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":72.2000000000000028421709430404007434844970703125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"phi4-14b","model_name":"Phi-4 14B Instruct","params_b":14.699999999999999289457264239899814128875732421875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":50.2000000000000028421709430404007434844970703125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"deepseek-r1-distill-qwen-7b","model_name":"DeepSeek-R1-Distill-Qwen-7B","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":65.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"deepseek-r1-distill-qwen-14b","model_name":"DeepSeek-R1-Distill-Qwen-14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":51.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"deepseek-r1-distill-qwen-32b","model_name":"DeepSeek-R1-Distill-Qwen-32B","params_b":32,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":15,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"deepseek-r1-distill-llama-70b","model_name":"DeepSeek-R1-Distill-Llama-70B","params_b":70,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":37,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"olmo2-7b","model_name":"OLMo 2 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":65.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"olmo2-13b","model_name":"OLMo 2 13B Instruct","params_b":13,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":53.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"command-r-35b","model_name":"Command R 35B (08-2024)","params_b":35,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":44.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"h100-sxm-80gb","gpu_name":"NVIDIA H100 SXM 80 GB","gpu_vram_gb":80,"model_id":"command-r-plus-104b","model_name":"Command R+ 104B","params_b":104,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":21,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"smollm2-1.7b","model_name":"SmolLM2 1.7B Instruct","params_b":1.6999999999999999555910790149937383830547332763671875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":76.400000000000005684341886080801486968994140625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"llama3.2-1b","model_name":"Llama 3.2 1B Instruct","params_b":1,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":77.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"llama3.2-3b","model_name":"Llama 3.2 3B Instruct","params_b":3,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":73.7000000000000028421709430404007434844970703125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"llama3.1-8b","model_name":"Llama 3.1 8B Instruct","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":63.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"llama3.1-70b","model_name":"Llama 3.1 70B Instruct","params_b":70,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":37,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"llama3.3-70b","model_name":"Llama 3.3 70B Instruct","params_b":70,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":37,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"qwen2.5-7b","model_name":"Qwen2.5 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":65.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"qwen2.5-14b","model_name":"Qwen2.5 14B Instruct","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":51.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"qwen2.5-32b","model_name":"Qwen2.5 32B Instruct","params_b":32,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":15,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"qwen2.5-72b","model_name":"Qwen2.5 72B Instruct","params_b":72,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":36,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"qwen3-8b","model_name":"Qwen3 8B","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":63.60000000000000142108547152020037174224853515625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"qwen3-14b","model_name":"Qwen3 14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":51.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"qwen3-32b","model_name":"Qwen3 32B","params_b":32,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":15,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"qwen3-30b-a3b","model_name":"Qwen3 30B-A3B (MoE)","params_b":30,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":18,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"qwen3-235b-a22b","model_name":"Qwen3 235B-A22B (MoE)","params_b":235,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"gemma3-4b","model_name":"Gemma 3 4B Instruct","params_b":4,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":71.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"gemma3-12b","model_name":"Gemma 3 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":55.60000000000000142108547152020037174224853515625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"gemma3-27b","model_name":"Gemma 3 27B Instruct","params_b":27,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":25.300000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"gemma4-e4b","model_name":"Gemma 4 E4B (Edge)","params_b":4.5,"q4_fits":true,"q4_quality":"ample","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"q4","headroom_at_recommended":70.400000000000005684341886080801486968994140625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"gemma4-31b","model_name":"Gemma 4 31B","params_b":31,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":17,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"mistral-7b-v0.3","model_name":"Mistral 7B v0.3 Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":65.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"mistral-nemo-12b","model_name":"Mistral Nemo 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":55.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"mixtral-8x7b","model_name":"Mixtral 8x7B Instruct v0.1","params_b":46.7000000000000028421709430404007434844970703125,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":32.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"phi3.5-mini-3.8b","model_name":"Phi-3.5 mini 3.8B Instruct","params_b":3.79999999999999982236431605997495353221893310546875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":72.2000000000000028421709430404007434844970703125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"phi4-14b","model_name":"Phi-4 14B Instruct","params_b":14.699999999999999289457264239899814128875732421875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":50.2000000000000028421709430404007434844970703125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"deepseek-r1-distill-qwen-7b","model_name":"DeepSeek-R1-Distill-Qwen-7B","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":65.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"deepseek-r1-distill-qwen-14b","model_name":"DeepSeek-R1-Distill-Qwen-14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":51.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"deepseek-r1-distill-qwen-32b","model_name":"DeepSeek-R1-Distill-Qwen-32B","params_b":32,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":15,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"deepseek-r1-distill-llama-70b","model_name":"DeepSeek-R1-Distill-Llama-70B","params_b":70,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":37,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"olmo2-7b","model_name":"OLMo 2 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":65.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"olmo2-13b","model_name":"OLMo 2 13B Instruct","params_b":13,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":53.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"command-r-35b","model_name":"Command R 35B (08-2024)","params_b":35,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":44.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"a100-sxm4-80gb","gpu_name":"NVIDIA A100 SXM4 80 GB","gpu_vram_gb":80,"model_id":"command-r-plus-104b","model_name":"Command R+ 104B","params_b":104,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":21,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"smollm2-1.7b","model_name":"SmolLM2 1.7B Instruct","params_b":1.6999999999999999555910790149937383830547332763671875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":44.39999999999999857891452847979962825775146484375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"llama3.2-1b","model_name":"Llama 3.2 1B Instruct","params_b":1,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":45.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"llama3.2-3b","model_name":"Llama 3.2 3B Instruct","params_b":3,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":41.7000000000000028421709430404007434844970703125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"llama3.1-8b","model_name":"Llama 3.1 8B Instruct","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":31.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"llama3.1-70b","model_name":"Llama 3.1 70B Instruct","params_b":70,"q4_fits":true,"q4_quality":"tight","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"llama3.3-70b","model_name":"Llama 3.3 70B Instruct","params_b":70,"q4_fits":true,"q4_quality":"tight","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"qwen2.5-7b","model_name":"Qwen2.5 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":33.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"qwen2.5-14b","model_name":"Qwen2.5 14B Instruct","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":19.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"qwen2.5-32b","model_name":"Qwen2.5 32B Instruct","params_b":32,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":15,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"qwen2.5-72b","model_name":"Qwen2.5 72B Instruct","params_b":72,"q4_fits":true,"q4_quality":"tight","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":4,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"qwen3-8b","model_name":"Qwen3 8B","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":31.60000000000000142108547152020037174224853515625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"qwen3-14b","model_name":"Qwen3 14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":19.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"qwen3-32b","model_name":"Qwen3 32B","params_b":32,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":15,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"qwen3-30b-a3b","model_name":"Qwen3 30B-A3B (MoE)","params_b":30,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":16,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"qwen3-235b-a22b","model_name":"Qwen3 235B-A22B (MoE)","params_b":235,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"gemma3-4b","model_name":"Gemma 3 4B Instruct","params_b":4,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":39.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"gemma3-12b","model_name":"Gemma 3 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":23.60000000000000142108547152020037174224853515625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"gemma3-27b","model_name":"Gemma 3 27B Instruct","params_b":27,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":20.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"gemma4-e4b","model_name":"Gemma 4 E4B (Edge)","params_b":4.5,"q4_fits":true,"q4_quality":"ample","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"q4","headroom_at_recommended":38.39999999999999857891452847979962825775146484375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"gemma4-31b","model_name":"Gemma 4 31B","params_b":31,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":15.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"mistral-7b-v0.3","model_name":"Mistral 7B v0.3 Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":33.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"mistral-nemo-12b","model_name":"Mistral Nemo 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":23.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"mixtral-8x7b","model_name":"Mixtral 8x7B Instruct v0.1","params_b":46.7000000000000028421709430404007434844970703125,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":22,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"phi3.5-mini-3.8b","model_name":"Phi-3.5 mini 3.8B Instruct","params_b":3.79999999999999982236431605997495353221893310546875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":40.2000000000000028421709430404007434844970703125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"phi4-14b","model_name":"Phi-4 14B Instruct","params_b":14.699999999999999289457264239899814128875732421875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":18.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"deepseek-r1-distill-qwen-7b","model_name":"DeepSeek-R1-Distill-Qwen-7B","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":33.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"deepseek-r1-distill-qwen-14b","model_name":"DeepSeek-R1-Distill-Qwen-14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":19.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"deepseek-r1-distill-qwen-32b","model_name":"DeepSeek-R1-Distill-Qwen-32B","params_b":32,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":15,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"deepseek-r1-distill-llama-70b","model_name":"DeepSeek-R1-Distill-Llama-70B","params_b":70,"q4_fits":true,"q4_quality":"tight","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"olmo2-7b","model_name":"OLMo 2 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":33.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"olmo2-13b","model_name":"OLMo 2 13B Instruct","params_b":13,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":21.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"command-r-35b","model_name":"Command R 35B (08-2024)","params_b":35,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":12.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l40s-48gb","gpu_name":"NVIDIA L40S 48 GB","gpu_vram_gb":48,"model_id":"command-r-plus-104b","model_name":"Command R+ 104B","params_b":104,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"smollm2-1.7b","model_name":"SmolLM2 1.7B Instruct","params_b":1.6999999999999999555910790149937383830547332763671875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":20.39999999999999857891452847979962825775146484375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"llama3.2-1b","model_name":"Llama 3.2 1B Instruct","params_b":1,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":21.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"llama3.2-3b","model_name":"Llama 3.2 3B Instruct","params_b":3,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":17.699999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"llama3.1-8b","model_name":"Llama 3.1 8B Instruct","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":7.79999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"llama3.1-70b","model_name":"Llama 3.1 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"llama3.3-70b","model_name":"Llama 3.3 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"qwen2.5-7b","model_name":"Qwen2.5 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":9.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"qwen2.5-14b","model_name":"Qwen2.5 14B Instruct","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":9.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"qwen2.5-32b","model_name":"Qwen2.5 32B Instruct","params_b":32,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"qwen2.5-72b","model_name":"Qwen2.5 72B Instruct","params_b":72,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"qwen3-8b","model_name":"Qwen3 8B","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":7.5999999999999996447286321199499070644378662109375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"qwen3-14b","model_name":"Qwen3 14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":9.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"qwen3-32b","model_name":"Qwen3 32B","params_b":32,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":4,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"qwen3-30b-a3b","model_name":"Qwen3 30B-A3B (MoE)","params_b":30,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"qwen3-235b-a22b","model_name":"Qwen3 235B-A22B (MoE)","params_b":235,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"gemma3-4b","model_name":"Gemma 3 4B Instruct","params_b":4,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":15.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"gemma3-12b","model_name":"Gemma 3 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":11.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"gemma3-27b","model_name":"Gemma 3 27B Instruct","params_b":27,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":7.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"gemma4-e4b","model_name":"Gemma 4 E4B (Edge)","params_b":4.5,"q4_fits":true,"q4_quality":"ample","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"q4","headroom_at_recommended":14.4000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"gemma4-31b","model_name":"Gemma 4 31B","params_b":31,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":4,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"mistral-7b-v0.3","model_name":"Mistral 7B v0.3 Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":9.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"mistral-nemo-12b","model_name":"Mistral Nemo 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":11.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"mixtral-8x7b","model_name":"Mixtral 8x7B Instruct v0.1","params_b":46.7000000000000028421709430404007434844970703125,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"phi3.5-mini-3.8b","model_name":"Phi-3.5 mini 3.8B Instruct","params_b":3.79999999999999982236431605997495353221893310546875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":16.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"phi4-14b","model_name":"Phi-4 14B Instruct","params_b":14.699999999999999289457264239899814128875732421875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"deepseek-r1-distill-qwen-7b","model_name":"DeepSeek-R1-Distill-Qwen-7B","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":9.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"deepseek-r1-distill-qwen-14b","model_name":"DeepSeek-R1-Distill-Qwen-14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":9.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"deepseek-r1-distill-qwen-32b","model_name":"DeepSeek-R1-Distill-Qwen-32B","params_b":32,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":4,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"deepseek-r1-distill-llama-70b","model_name":"DeepSeek-R1-Distill-Llama-70B","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"olmo2-7b","model_name":"OLMo 2 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":9.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"olmo2-13b","model_name":"OLMo 2 13B Instruct","params_b":13,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":10.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"command-r-35b","model_name":"Command R 35B (08-2024)","params_b":35,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"l4-24gb","gpu_name":"NVIDIA L4 24 GB","gpu_vram_gb":24,"model_id":"command-r-plus-104b","model_name":"Command R+ 104B","params_b":104,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"smollm2-1.7b","model_name":"SmolLM2 1.7B Instruct","params_b":1.6999999999999999555910790149937383830547332763671875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":44.39999999999999857891452847979962825775146484375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"llama3.2-1b","model_name":"Llama 3.2 1B Instruct","params_b":1,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":45.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"llama3.2-3b","model_name":"Llama 3.2 3B Instruct","params_b":3,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":41.7000000000000028421709430404007434844970703125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"llama3.1-8b","model_name":"Llama 3.1 8B Instruct","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":31.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"llama3.1-70b","model_name":"Llama 3.1 70B Instruct","params_b":70,"q4_fits":true,"q4_quality":"tight","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"llama3.3-70b","model_name":"Llama 3.3 70B Instruct","params_b":70,"q4_fits":true,"q4_quality":"tight","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"qwen2.5-7b","model_name":"Qwen2.5 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":33.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"qwen2.5-14b","model_name":"Qwen2.5 14B Instruct","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":19.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"qwen2.5-32b","model_name":"Qwen2.5 32B Instruct","params_b":32,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":15,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"qwen2.5-72b","model_name":"Qwen2.5 72B Instruct","params_b":72,"q4_fits":true,"q4_quality":"tight","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":4,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"qwen3-8b","model_name":"Qwen3 8B","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":31.60000000000000142108547152020037174224853515625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"qwen3-14b","model_name":"Qwen3 14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":19.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"qwen3-32b","model_name":"Qwen3 32B","params_b":32,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":15,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"qwen3-30b-a3b","model_name":"Qwen3 30B-A3B (MoE)","params_b":30,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":16,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"qwen3-235b-a22b","model_name":"Qwen3 235B-A22B (MoE)","params_b":235,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"gemma3-4b","model_name":"Gemma 3 4B Instruct","params_b":4,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":39.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"gemma3-12b","model_name":"Gemma 3 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":23.60000000000000142108547152020037174224853515625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"gemma3-27b","model_name":"Gemma 3 27B Instruct","params_b":27,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":20.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"gemma4-e4b","model_name":"Gemma 4 E4B (Edge)","params_b":4.5,"q4_fits":true,"q4_quality":"ample","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"q4","headroom_at_recommended":38.39999999999999857891452847979962825775146484375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"gemma4-31b","model_name":"Gemma 4 31B","params_b":31,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":15.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"mistral-7b-v0.3","model_name":"Mistral 7B v0.3 Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":33.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"mistral-nemo-12b","model_name":"Mistral Nemo 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":23.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"mixtral-8x7b","model_name":"Mixtral 8x7B Instruct v0.1","params_b":46.7000000000000028421709430404007434844970703125,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":22,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"phi3.5-mini-3.8b","model_name":"Phi-3.5 mini 3.8B Instruct","params_b":3.79999999999999982236431605997495353221893310546875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":40.2000000000000028421709430404007434844970703125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"phi4-14b","model_name":"Phi-4 14B Instruct","params_b":14.699999999999999289457264239899814128875732421875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":18.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"deepseek-r1-distill-qwen-7b","model_name":"DeepSeek-R1-Distill-Qwen-7B","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":33.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"deepseek-r1-distill-qwen-14b","model_name":"DeepSeek-R1-Distill-Qwen-14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":19.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"deepseek-r1-distill-qwen-32b","model_name":"DeepSeek-R1-Distill-Qwen-32B","params_b":32,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":15,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"deepseek-r1-distill-llama-70b","model_name":"DeepSeek-R1-Distill-Llama-70B","params_b":70,"q4_fits":true,"q4_quality":"tight","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"olmo2-7b","model_name":"OLMo 2 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":33.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"olmo2-13b","model_name":"OLMo 2 13B Instruct","params_b":13,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":21.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"command-r-35b","model_name":"Command R 35B (08-2024)","params_b":35,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":12.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-a6000-ampere-48gb","gpu_name":"NVIDIA RTX A6000 (Ampere) 48 GB","gpu_vram_gb":48,"model_id":"command-r-plus-104b","model_name":"Command R+ 104B","params_b":104,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"smollm2-1.7b","model_name":"SmolLM2 1.7B Instruct","params_b":1.6999999999999999555910790149937383830547332763671875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":28.39999999999999857891452847979962825775146484375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"llama3.2-1b","model_name":"Llama 3.2 1B Instruct","params_b":1,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":29.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"llama3.2-3b","model_name":"Llama 3.2 3B Instruct","params_b":3,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":25.699999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"llama3.1-8b","model_name":"Llama 3.1 8B Instruct","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":15.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"llama3.1-70b","model_name":"Llama 3.1 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"llama3.3-70b","model_name":"Llama 3.3 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"qwen2.5-7b","model_name":"Qwen2.5 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":17.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"qwen2.5-14b","model_name":"Qwen2.5 14B Instruct","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":17.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"qwen2.5-32b","model_name":"Qwen2.5 32B Instruct","params_b":32,"q4_fits":true,"q4_quality":"ample","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":13,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"qwen2.5-72b","model_name":"Qwen2.5 72B Instruct","params_b":72,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"qwen3-8b","model_name":"Qwen3 8B","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":15.5999999999999996447286321199499070644378662109375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"qwen3-14b","model_name":"Qwen3 14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":17.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"qwen3-32b","model_name":"Qwen3 32B","params_b":32,"q4_fits":true,"q4_quality":"ample","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":12,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"qwen3-30b-a3b","model_name":"Qwen3 30B-A3B (MoE)","params_b":30,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":13,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"qwen3-235b-a22b","model_name":"Qwen3 235B-A22B (MoE)","params_b":235,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"gemma3-4b","model_name":"Gemma 3 4B Instruct","params_b":4,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":23.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"gemma3-12b","model_name":"Gemma 3 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":7.5999999999999996447286321199499070644378662109375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"gemma3-27b","model_name":"Gemma 3 27B Instruct","params_b":27,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":4.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"gemma4-e4b","model_name":"Gemma 4 E4B (Edge)","params_b":4.5,"q4_fits":true,"q4_quality":"ample","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"q4","headroom_at_recommended":22.39999999999999857891452847979962825775146484375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"gemma4-31b","model_name":"Gemma 4 31B","params_b":31,"q4_fits":true,"q4_quality":"ample","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":12,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"mistral-7b-v0.3","model_name":"Mistral 7B v0.3 Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":17.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"mistral-nemo-12b","model_name":"Mistral Nemo 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":7.79999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"mixtral-8x7b","model_name":"Mixtral 8x7B Instruct v0.1","params_b":46.7000000000000028421709430404007434844970703125,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":6,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"phi3.5-mini-3.8b","model_name":"Phi-3.5 mini 3.8B Instruct","params_b":3.79999999999999982236431605997495353221893310546875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":24.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"phi4-14b","model_name":"Phi-4 14B Instruct","params_b":14.699999999999999289457264239899814128875732421875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":16.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"deepseek-r1-distill-qwen-7b","model_name":"DeepSeek-R1-Distill-Qwen-7B","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":17.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"deepseek-r1-distill-qwen-14b","model_name":"DeepSeek-R1-Distill-Qwen-14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":17.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"deepseek-r1-distill-qwen-32b","model_name":"DeepSeek-R1-Distill-Qwen-32B","params_b":32,"q4_fits":true,"q4_quality":"ample","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":12,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"deepseek-r1-distill-llama-70b","model_name":"DeepSeek-R1-Distill-Llama-70B","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"olmo2-7b","model_name":"OLMo 2 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":17.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"olmo2-13b","model_name":"OLMo 2 13B Instruct","params_b":13,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":5.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"command-r-35b","model_name":"Command R 35B (08-2024)","params_b":35,"q4_fits":true,"q4_quality":"ample","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":13,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5090-32gb","gpu_name":"NVIDIA GeForce RTX 5090","gpu_vram_gb":32,"model_id":"command-r-plus-104b","model_name":"Command R+ 104B","params_b":104,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"smollm2-1.7b","model_name":"SmolLM2 1.7B Instruct","params_b":1.6999999999999999555910790149937383830547332763671875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":20.39999999999999857891452847979962825775146484375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"llama3.2-1b","model_name":"Llama 3.2 1B Instruct","params_b":1,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":21.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"llama3.2-3b","model_name":"Llama 3.2 3B Instruct","params_b":3,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":17.699999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"llama3.1-8b","model_name":"Llama 3.1 8B Instruct","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":7.79999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"llama3.1-70b","model_name":"Llama 3.1 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"llama3.3-70b","model_name":"Llama 3.3 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"qwen2.5-7b","model_name":"Qwen2.5 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":9.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"qwen2.5-14b","model_name":"Qwen2.5 14B Instruct","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":9.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"qwen2.5-32b","model_name":"Qwen2.5 32B Instruct","params_b":32,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"qwen2.5-72b","model_name":"Qwen2.5 72B Instruct","params_b":72,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"qwen3-8b","model_name":"Qwen3 8B","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":7.5999999999999996447286321199499070644378662109375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"qwen3-14b","model_name":"Qwen3 14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":9.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"qwen3-32b","model_name":"Qwen3 32B","params_b":32,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":4,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"qwen3-30b-a3b","model_name":"Qwen3 30B-A3B (MoE)","params_b":30,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"qwen3-235b-a22b","model_name":"Qwen3 235B-A22B (MoE)","params_b":235,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"gemma3-4b","model_name":"Gemma 3 4B Instruct","params_b":4,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":15.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"gemma3-12b","model_name":"Gemma 3 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":11.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"gemma3-27b","model_name":"Gemma 3 27B Instruct","params_b":27,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":7.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"gemma4-e4b","model_name":"Gemma 4 E4B (Edge)","params_b":4.5,"q4_fits":true,"q4_quality":"ample","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"q4","headroom_at_recommended":14.4000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"gemma4-31b","model_name":"Gemma 4 31B","params_b":31,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":4,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"mistral-7b-v0.3","model_name":"Mistral 7B v0.3 Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":9.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"mistral-nemo-12b","model_name":"Mistral Nemo 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":11.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"mixtral-8x7b","model_name":"Mixtral 8x7B Instruct v0.1","params_b":46.7000000000000028421709430404007434844970703125,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"phi3.5-mini-3.8b","model_name":"Phi-3.5 mini 3.8B Instruct","params_b":3.79999999999999982236431605997495353221893310546875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":16.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"phi4-14b","model_name":"Phi-4 14B Instruct","params_b":14.699999999999999289457264239899814128875732421875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"deepseek-r1-distill-qwen-7b","model_name":"DeepSeek-R1-Distill-Qwen-7B","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":9.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"deepseek-r1-distill-qwen-14b","model_name":"DeepSeek-R1-Distill-Qwen-14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":9.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"deepseek-r1-distill-qwen-32b","model_name":"DeepSeek-R1-Distill-Qwen-32B","params_b":32,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":4,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"deepseek-r1-distill-llama-70b","model_name":"DeepSeek-R1-Distill-Llama-70B","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"olmo2-7b","model_name":"OLMo 2 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":9.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"olmo2-13b","model_name":"OLMo 2 13B Instruct","params_b":13,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":10.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"command-r-35b","model_name":"Command R 35B (08-2024)","params_b":35,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4090-24gb","gpu_name":"NVIDIA GeForce RTX 4090","gpu_vram_gb":24,"model_id":"command-r-plus-104b","model_name":"Command R+ 104B","params_b":104,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"smollm2-1.7b","model_name":"SmolLM2 1.7B Instruct","params_b":1.6999999999999999555910790149937383830547332763671875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":124.400000000000005684341886080801486968994140625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"llama3.2-1b","model_name":"Llama 3.2 1B Instruct","params_b":1,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":125.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"llama3.2-3b","model_name":"Llama 3.2 3B Instruct","params_b":3,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":121.7000000000000028421709430404007434844970703125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"llama3.1-8b","model_name":"Llama 3.1 8B Instruct","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":111.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"llama3.1-70b","model_name":"Llama 3.1 70B Instruct","params_b":70,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":53,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"llama3.3-70b","model_name":"Llama 3.3 70B Instruct","params_b":70,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":53,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"qwen2.5-7b","model_name":"Qwen2.5 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":113.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"qwen2.5-14b","model_name":"Qwen2.5 14B Instruct","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":99.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"qwen2.5-32b","model_name":"Qwen2.5 32B Instruct","params_b":32,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":63,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"qwen2.5-72b","model_name":"Qwen2.5 72B Instruct","params_b":72,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":53,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"qwen3-8b","model_name":"Qwen3 8B","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":111.599999999999994315658113919198513031005859375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"qwen3-14b","model_name":"Qwen3 14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":99.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"qwen3-32b","model_name":"Qwen3 32B","params_b":32,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":63,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"qwen3-30b-a3b","model_name":"Qwen3 30B-A3B (MoE)","params_b":30,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":66,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"qwen3-235b-a22b","model_name":"Qwen3 235B-A22B (MoE)","params_b":235,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"gemma3-4b","model_name":"Gemma 3 4B Instruct","params_b":4,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":119.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"gemma3-12b","model_name":"Gemma 3 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":103.599999999999994315658113919198513031005859375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"gemma3-27b","model_name":"Gemma 3 27B Instruct","params_b":27,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":73.2999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"gemma4-e4b","model_name":"Gemma 4 E4B (Edge)","params_b":4.5,"q4_fits":true,"q4_quality":"ample","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"q4","headroom_at_recommended":118.400000000000005684341886080801486968994140625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"gemma4-31b","model_name":"Gemma 4 31B","params_b":31,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":65,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"mistral-7b-v0.3","model_name":"Mistral 7B v0.3 Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":113.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"mistral-nemo-12b","model_name":"Mistral Nemo 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":103.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"mixtral-8x7b","model_name":"Mixtral 8x7B Instruct v0.1","params_b":46.7000000000000028421709430404007434844970703125,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":34.2000000000000028421709430404007434844970703125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"phi3.5-mini-3.8b","model_name":"Phi-3.5 mini 3.8B Instruct","params_b":3.79999999999999982236431605997495353221893310546875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":120.2000000000000028421709430404007434844970703125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"phi4-14b","model_name":"Phi-4 14B Instruct","params_b":14.699999999999999289457264239899814128875732421875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":98.2000000000000028421709430404007434844970703125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"deepseek-r1-distill-qwen-7b","model_name":"DeepSeek-R1-Distill-Qwen-7B","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":113.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"deepseek-r1-distill-qwen-14b","model_name":"DeepSeek-R1-Distill-Qwen-14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":99.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"deepseek-r1-distill-qwen-32b","model_name":"DeepSeek-R1-Distill-Qwen-32B","params_b":32,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":63,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"deepseek-r1-distill-llama-70b","model_name":"DeepSeek-R1-Distill-Llama-70B","params_b":70,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":53,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"olmo2-7b","model_name":"OLMo 2 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":113.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"olmo2-13b","model_name":"OLMo 2 13B Instruct","params_b":13,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":101.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"command-r-35b","model_name":"Command R 35B (08-2024)","params_b":35,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":57.10000000000000142108547152020037174224853515625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-max-128gb","gpu_name":"Apple M4 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"command-r-plus-104b","model_name":"Command R+ 104B","params_b":104,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":18,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"smollm2-1.7b","model_name":"SmolLM2 1.7B Instruct","params_b":1.6999999999999999555910790149937383830547332763671875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":20.39999999999999857891452847979962825775146484375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"llama3.2-1b","model_name":"Llama 3.2 1B Instruct","params_b":1,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":21.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"llama3.2-3b","model_name":"Llama 3.2 3B Instruct","params_b":3,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":17.699999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"llama3.1-8b","model_name":"Llama 3.1 8B Instruct","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":7.79999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"llama3.1-70b","model_name":"Llama 3.1 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"llama3.3-70b","model_name":"Llama 3.3 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"qwen2.5-7b","model_name":"Qwen2.5 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":9.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"qwen2.5-14b","model_name":"Qwen2.5 14B Instruct","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":9.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"qwen2.5-32b","model_name":"Qwen2.5 32B Instruct","params_b":32,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"qwen2.5-72b","model_name":"Qwen2.5 72B Instruct","params_b":72,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"qwen3-8b","model_name":"Qwen3 8B","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":7.5999999999999996447286321199499070644378662109375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"qwen3-14b","model_name":"Qwen3 14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":9.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"qwen3-32b","model_name":"Qwen3 32B","params_b":32,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":4,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"qwen3-30b-a3b","model_name":"Qwen3 30B-A3B (MoE)","params_b":30,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"qwen3-235b-a22b","model_name":"Qwen3 235B-A22B (MoE)","params_b":235,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"gemma3-4b","model_name":"Gemma 3 4B Instruct","params_b":4,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":15.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"gemma3-12b","model_name":"Gemma 3 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":11.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"gemma3-27b","model_name":"Gemma 3 27B Instruct","params_b":27,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":7.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"gemma4-e4b","model_name":"Gemma 4 E4B (Edge)","params_b":4.5,"q4_fits":true,"q4_quality":"ample","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"q4","headroom_at_recommended":14.4000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"gemma4-31b","model_name":"Gemma 4 31B","params_b":31,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":4,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"mistral-7b-v0.3","model_name":"Mistral 7B v0.3 Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":9.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"mistral-nemo-12b","model_name":"Mistral Nemo 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":11.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"mixtral-8x7b","model_name":"Mixtral 8x7B Instruct v0.1","params_b":46.7000000000000028421709430404007434844970703125,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"phi3.5-mini-3.8b","model_name":"Phi-3.5 mini 3.8B Instruct","params_b":3.79999999999999982236431605997495353221893310546875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":16.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"phi4-14b","model_name":"Phi-4 14B Instruct","params_b":14.699999999999999289457264239899814128875732421875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"deepseek-r1-distill-qwen-7b","model_name":"DeepSeek-R1-Distill-Qwen-7B","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":9.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"deepseek-r1-distill-qwen-14b","model_name":"DeepSeek-R1-Distill-Qwen-14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":9.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"deepseek-r1-distill-qwen-32b","model_name":"DeepSeek-R1-Distill-Qwen-32B","params_b":32,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":4,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"deepseek-r1-distill-llama-70b","model_name":"DeepSeek-R1-Distill-Llama-70B","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"olmo2-7b","model_name":"OLMo 2 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":9.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"olmo2-13b","model_name":"OLMo 2 13B Instruct","params_b":13,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":10.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"command-r-35b","model_name":"Command R 35B (08-2024)","params_b":35,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3090-24gb","gpu_name":"NVIDIA GeForce RTX 3090","gpu_vram_gb":24,"model_id":"command-r-plus-104b","model_name":"Command R+ 104B","params_b":104,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"smollm2-1.7b","model_name":"SmolLM2 1.7B Instruct","params_b":1.6999999999999999555910790149937383830547332763671875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":12.4000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"llama3.2-1b","model_name":"Llama 3.2 1B Instruct","params_b":1,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":13.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"llama3.2-3b","model_name":"Llama 3.2 3B Instruct","params_b":3,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":9.699999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"llama3.1-8b","model_name":"Llama 3.1 8B Instruct","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":7.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"llama3.1-70b","model_name":"Llama 3.1 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"llama3.3-70b","model_name":"Llama 3.3 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"qwen2.5-7b","model_name":"Qwen2.5 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"qwen2.5-14b","model_name":"Qwen2.5 14B Instruct","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":7.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"qwen2.5-32b","model_name":"Qwen2.5 32B Instruct","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"qwen2.5-72b","model_name":"Qwen2.5 72B Instruct","params_b":72,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"qwen3-8b","model_name":"Qwen3 8B","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":7.20000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"qwen3-14b","model_name":"Qwen3 14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":6.70000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"qwen3-32b","model_name":"Qwen3 32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"qwen3-30b-a3b","model_name":"Qwen3 30B-A3B (MoE)","params_b":30,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"qwen3-235b-a22b","model_name":"Qwen3 235B-A22B (MoE)","params_b":235,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"gemma3-4b","model_name":"Gemma 3 4B Instruct","params_b":4,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":7.79999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"gemma3-12b","model_name":"Gemma 3 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":3.20000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"gemma3-27b","model_name":"Gemma 3 27B Instruct","params_b":27,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"gemma4-e4b","model_name":"Gemma 4 E4B (Edge)","params_b":4.5,"q4_fits":true,"q4_quality":"ample","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"q4","headroom_at_recommended":6.4000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"gemma4-31b","model_name":"Gemma 4 31B","params_b":31,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"mistral-7b-v0.3","model_name":"Mistral 7B v0.3 Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"mistral-nemo-12b","model_name":"Mistral Nemo 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":3.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"mixtral-8x7b","model_name":"Mixtral 8x7B Instruct v0.1","params_b":46.7000000000000028421709430404007434844970703125,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"phi3.5-mini-3.8b","model_name":"Phi-3.5 mini 3.8B Instruct","params_b":3.79999999999999982236431605997495353221893310546875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":8.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"phi4-14b","model_name":"Phi-4 14B Instruct","params_b":14.699999999999999289457264239899814128875732421875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":6.9000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"deepseek-r1-distill-qwen-7b","model_name":"DeepSeek-R1-Distill-Qwen-7B","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"deepseek-r1-distill-qwen-14b","model_name":"DeepSeek-R1-Distill-Qwen-14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":7,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"deepseek-r1-distill-qwen-32b","model_name":"DeepSeek-R1-Distill-Qwen-32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"deepseek-r1-distill-llama-70b","model_name":"DeepSeek-R1-Distill-Llama-70B","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"olmo2-7b","model_name":"OLMo 2 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"olmo2-13b","model_name":"OLMo 2 13B Instruct","params_b":13,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":2.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"command-r-35b","model_name":"Command R 35B (08-2024)","params_b":35,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-5080-16gb","gpu_name":"NVIDIA GeForce RTX 5080","gpu_vram_gb":16,"model_id":"command-r-plus-104b","model_name":"Command R+ 104B","params_b":104,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"smollm2-1.7b","model_name":"SmolLM2 1.7B Instruct","params_b":1.6999999999999999555910790149937383830547332763671875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":12.4000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"llama3.2-1b","model_name":"Llama 3.2 1B Instruct","params_b":1,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":13.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"llama3.2-3b","model_name":"Llama 3.2 3B Instruct","params_b":3,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":9.699999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"llama3.1-8b","model_name":"Llama 3.1 8B Instruct","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":7.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"llama3.1-70b","model_name":"Llama 3.1 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"llama3.3-70b","model_name":"Llama 3.3 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"qwen2.5-7b","model_name":"Qwen2.5 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"qwen2.5-14b","model_name":"Qwen2.5 14B Instruct","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":7.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"qwen2.5-32b","model_name":"Qwen2.5 32B Instruct","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"qwen2.5-72b","model_name":"Qwen2.5 72B Instruct","params_b":72,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"qwen3-8b","model_name":"Qwen3 8B","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":7.20000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"qwen3-14b","model_name":"Qwen3 14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":6.70000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"qwen3-32b","model_name":"Qwen3 32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"qwen3-30b-a3b","model_name":"Qwen3 30B-A3B (MoE)","params_b":30,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"qwen3-235b-a22b","model_name":"Qwen3 235B-A22B (MoE)","params_b":235,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"gemma3-4b","model_name":"Gemma 3 4B Instruct","params_b":4,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":7.79999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"gemma3-12b","model_name":"Gemma 3 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":3.20000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"gemma3-27b","model_name":"Gemma 3 27B Instruct","params_b":27,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"gemma4-e4b","model_name":"Gemma 4 E4B (Edge)","params_b":4.5,"q4_fits":true,"q4_quality":"ample","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"q4","headroom_at_recommended":6.4000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"gemma4-31b","model_name":"Gemma 4 31B","params_b":31,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"mistral-7b-v0.3","model_name":"Mistral 7B v0.3 Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"mistral-nemo-12b","model_name":"Mistral Nemo 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":3.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"mixtral-8x7b","model_name":"Mixtral 8x7B Instruct v0.1","params_b":46.7000000000000028421709430404007434844970703125,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"phi3.5-mini-3.8b","model_name":"Phi-3.5 mini 3.8B Instruct","params_b":3.79999999999999982236431605997495353221893310546875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":8.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"phi4-14b","model_name":"Phi-4 14B Instruct","params_b":14.699999999999999289457264239899814128875732421875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":6.9000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"deepseek-r1-distill-qwen-7b","model_name":"DeepSeek-R1-Distill-Qwen-7B","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"deepseek-r1-distill-qwen-14b","model_name":"DeepSeek-R1-Distill-Qwen-14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":7,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"deepseek-r1-distill-qwen-32b","model_name":"DeepSeek-R1-Distill-Qwen-32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"deepseek-r1-distill-llama-70b","model_name":"DeepSeek-R1-Distill-Llama-70B","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"olmo2-7b","model_name":"OLMo 2 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"olmo2-13b","model_name":"OLMo 2 13B Instruct","params_b":13,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":2.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"command-r-35b","model_name":"Command R 35B (08-2024)","params_b":35,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4080-super-16gb","gpu_name":"NVIDIA GeForce RTX 4080 Super","gpu_vram_gb":16,"model_id":"command-r-plus-104b","model_name":"Command R+ 104B","params_b":104,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"smollm2-1.7b","model_name":"SmolLM2 1.7B Instruct","params_b":1.6999999999999999555910790149937383830547332763671875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":12.4000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"llama3.2-1b","model_name":"Llama 3.2 1B Instruct","params_b":1,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":13.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"llama3.2-3b","model_name":"Llama 3.2 3B Instruct","params_b":3,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":9.699999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"llama3.1-8b","model_name":"Llama 3.1 8B Instruct","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":7.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"llama3.1-70b","model_name":"Llama 3.1 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"llama3.3-70b","model_name":"Llama 3.3 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"qwen2.5-7b","model_name":"Qwen2.5 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"qwen2.5-14b","model_name":"Qwen2.5 14B Instruct","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":7.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"qwen2.5-32b","model_name":"Qwen2.5 32B Instruct","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"qwen2.5-72b","model_name":"Qwen2.5 72B Instruct","params_b":72,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"qwen3-8b","model_name":"Qwen3 8B","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":7.20000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"qwen3-14b","model_name":"Qwen3 14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":6.70000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"qwen3-32b","model_name":"Qwen3 32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"qwen3-30b-a3b","model_name":"Qwen3 30B-A3B (MoE)","params_b":30,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"qwen3-235b-a22b","model_name":"Qwen3 235B-A22B (MoE)","params_b":235,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"gemma3-4b","model_name":"Gemma 3 4B Instruct","params_b":4,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":7.79999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"gemma3-12b","model_name":"Gemma 3 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":3.20000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"gemma3-27b","model_name":"Gemma 3 27B Instruct","params_b":27,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"gemma4-e4b","model_name":"Gemma 4 E4B (Edge)","params_b":4.5,"q4_fits":true,"q4_quality":"ample","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"q4","headroom_at_recommended":6.4000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"gemma4-31b","model_name":"Gemma 4 31B","params_b":31,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"mistral-7b-v0.3","model_name":"Mistral 7B v0.3 Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"mistral-nemo-12b","model_name":"Mistral Nemo 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":3.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"mixtral-8x7b","model_name":"Mixtral 8x7B Instruct v0.1","params_b":46.7000000000000028421709430404007434844970703125,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"phi3.5-mini-3.8b","model_name":"Phi-3.5 mini 3.8B Instruct","params_b":3.79999999999999982236431605997495353221893310546875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":8.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"phi4-14b","model_name":"Phi-4 14B Instruct","params_b":14.699999999999999289457264239899814128875732421875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":6.9000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"deepseek-r1-distill-qwen-7b","model_name":"DeepSeek-R1-Distill-Qwen-7B","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"deepseek-r1-distill-qwen-14b","model_name":"DeepSeek-R1-Distill-Qwen-14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":7,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"deepseek-r1-distill-qwen-32b","model_name":"DeepSeek-R1-Distill-Qwen-32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"deepseek-r1-distill-llama-70b","model_name":"DeepSeek-R1-Distill-Llama-70B","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"olmo2-7b","model_name":"OLMo 2 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"olmo2-13b","model_name":"OLMo 2 13B Instruct","params_b":13,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":2.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"command-r-35b","model_name":"Command R 35B (08-2024)","params_b":35,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4080-16gb","gpu_name":"NVIDIA GeForce RTX 4080","gpu_vram_gb":16,"model_id":"command-r-plus-104b","model_name":"Command R+ 104B","params_b":104,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"smollm2-1.7b","model_name":"SmolLM2 1.7B Instruct","params_b":1.6999999999999999555910790149937383830547332763671875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":20.39999999999999857891452847979962825775146484375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"llama3.2-1b","model_name":"Llama 3.2 1B Instruct","params_b":1,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":21.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"llama3.2-3b","model_name":"Llama 3.2 3B Instruct","params_b":3,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":17.699999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"llama3.1-8b","model_name":"Llama 3.1 8B Instruct","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":7.79999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"llama3.1-70b","model_name":"Llama 3.1 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"llama3.3-70b","model_name":"Llama 3.3 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"qwen2.5-7b","model_name":"Qwen2.5 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":9.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"qwen2.5-14b","model_name":"Qwen2.5 14B Instruct","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":9.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"qwen2.5-32b","model_name":"Qwen2.5 32B Instruct","params_b":32,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"qwen2.5-72b","model_name":"Qwen2.5 72B Instruct","params_b":72,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"qwen3-8b","model_name":"Qwen3 8B","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":7.5999999999999996447286321199499070644378662109375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"qwen3-14b","model_name":"Qwen3 14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":9.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"qwen3-32b","model_name":"Qwen3 32B","params_b":32,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":4,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"qwen3-30b-a3b","model_name":"Qwen3 30B-A3B (MoE)","params_b":30,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"qwen3-235b-a22b","model_name":"Qwen3 235B-A22B (MoE)","params_b":235,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"gemma3-4b","model_name":"Gemma 3 4B Instruct","params_b":4,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":15.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"gemma3-12b","model_name":"Gemma 3 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":11.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"gemma3-27b","model_name":"Gemma 3 27B Instruct","params_b":27,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":7.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"gemma4-e4b","model_name":"Gemma 4 E4B (Edge)","params_b":4.5,"q4_fits":true,"q4_quality":"ample","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"q4","headroom_at_recommended":14.4000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"gemma4-31b","model_name":"Gemma 4 31B","params_b":31,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":4,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"mistral-7b-v0.3","model_name":"Mistral 7B v0.3 Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":9.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"mistral-nemo-12b","model_name":"Mistral Nemo 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":11.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"mixtral-8x7b","model_name":"Mixtral 8x7B Instruct v0.1","params_b":46.7000000000000028421709430404007434844970703125,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"phi3.5-mini-3.8b","model_name":"Phi-3.5 mini 3.8B Instruct","params_b":3.79999999999999982236431605997495353221893310546875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":16.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"phi4-14b","model_name":"Phi-4 14B Instruct","params_b":14.699999999999999289457264239899814128875732421875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"deepseek-r1-distill-qwen-7b","model_name":"DeepSeek-R1-Distill-Qwen-7B","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":9.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"deepseek-r1-distill-qwen-14b","model_name":"DeepSeek-R1-Distill-Qwen-14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":9.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"deepseek-r1-distill-qwen-32b","model_name":"DeepSeek-R1-Distill-Qwen-32B","params_b":32,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":4,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"deepseek-r1-distill-llama-70b","model_name":"DeepSeek-R1-Distill-Llama-70B","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"olmo2-7b","model_name":"OLMo 2 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":9.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"olmo2-13b","model_name":"OLMo 2 13B Instruct","params_b":13,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":10.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"command-r-35b","model_name":"Command R 35B (08-2024)","params_b":35,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xtx-24gb","gpu_name":"AMD Radeon RX 7900 XTX","gpu_vram_gb":24,"model_id":"command-r-plus-104b","model_name":"Command R+ 104B","params_b":104,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"smollm2-1.7b","model_name":"SmolLM2 1.7B Instruct","params_b":1.6999999999999999555910790149937383830547332763671875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":124.400000000000005684341886080801486968994140625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"llama3.2-1b","model_name":"Llama 3.2 1B Instruct","params_b":1,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":125.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"llama3.2-3b","model_name":"Llama 3.2 3B Instruct","params_b":3,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":121.7000000000000028421709430404007434844970703125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"llama3.1-8b","model_name":"Llama 3.1 8B Instruct","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":111.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"llama3.1-70b","model_name":"Llama 3.1 70B Instruct","params_b":70,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":53,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"llama3.3-70b","model_name":"Llama 3.3 70B Instruct","params_b":70,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":53,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"qwen2.5-7b","model_name":"Qwen2.5 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":113.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"qwen2.5-14b","model_name":"Qwen2.5 14B Instruct","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":99.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"qwen2.5-32b","model_name":"Qwen2.5 32B Instruct","params_b":32,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":63,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"qwen2.5-72b","model_name":"Qwen2.5 72B Instruct","params_b":72,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":53,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"qwen3-8b","model_name":"Qwen3 8B","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":111.599999999999994315658113919198513031005859375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"qwen3-14b","model_name":"Qwen3 14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":99.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"qwen3-32b","model_name":"Qwen3 32B","params_b":32,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":63,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"qwen3-30b-a3b","model_name":"Qwen3 30B-A3B (MoE)","params_b":30,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":66,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"qwen3-235b-a22b","model_name":"Qwen3 235B-A22B (MoE)","params_b":235,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"gemma3-4b","model_name":"Gemma 3 4B Instruct","params_b":4,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":119.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"gemma3-12b","model_name":"Gemma 3 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":103.599999999999994315658113919198513031005859375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"gemma3-27b","model_name":"Gemma 3 27B Instruct","params_b":27,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":73.2999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"gemma4-e4b","model_name":"Gemma 4 E4B (Edge)","params_b":4.5,"q4_fits":true,"q4_quality":"ample","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"q4","headroom_at_recommended":118.400000000000005684341886080801486968994140625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"gemma4-31b","model_name":"Gemma 4 31B","params_b":31,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":65,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"mistral-7b-v0.3","model_name":"Mistral 7B v0.3 Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":113.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"mistral-nemo-12b","model_name":"Mistral Nemo 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":103.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"mixtral-8x7b","model_name":"Mixtral 8x7B Instruct v0.1","params_b":46.7000000000000028421709430404007434844970703125,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":34.2000000000000028421709430404007434844970703125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"phi3.5-mini-3.8b","model_name":"Phi-3.5 mini 3.8B Instruct","params_b":3.79999999999999982236431605997495353221893310546875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":120.2000000000000028421709430404007434844970703125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"phi4-14b","model_name":"Phi-4 14B Instruct","params_b":14.699999999999999289457264239899814128875732421875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":98.2000000000000028421709430404007434844970703125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"deepseek-r1-distill-qwen-7b","model_name":"DeepSeek-R1-Distill-Qwen-7B","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":113.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"deepseek-r1-distill-qwen-14b","model_name":"DeepSeek-R1-Distill-Qwen-14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":99.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"deepseek-r1-distill-qwen-32b","model_name":"DeepSeek-R1-Distill-Qwen-32B","params_b":32,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":63,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"deepseek-r1-distill-llama-70b","model_name":"DeepSeek-R1-Distill-Llama-70B","params_b":70,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":53,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"olmo2-7b","model_name":"OLMo 2 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":113.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"olmo2-13b","model_name":"OLMo 2 13B Instruct","params_b":13,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":101.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"command-r-35b","model_name":"Command R 35B (08-2024)","params_b":35,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":57.10000000000000142108547152020037174224853515625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m3-max-128gb","gpu_name":"Apple M3 Max (40-core GPU, up to 128 GB)","gpu_vram_gb":128,"model_id":"command-r-plus-104b","model_name":"Command R+ 104B","params_b":104,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":18,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"smollm2-1.7b","model_name":"SmolLM2 1.7B Instruct","params_b":1.6999999999999999555910790149937383830547332763671875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":12.4000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"llama3.2-1b","model_name":"Llama 3.2 1B Instruct","params_b":1,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":13.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"llama3.2-3b","model_name":"Llama 3.2 3B Instruct","params_b":3,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":9.699999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"llama3.1-8b","model_name":"Llama 3.1 8B Instruct","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":7.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"llama3.1-70b","model_name":"Llama 3.1 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"llama3.3-70b","model_name":"Llama 3.3 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"qwen2.5-7b","model_name":"Qwen2.5 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"qwen2.5-14b","model_name":"Qwen2.5 14B Instruct","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":7.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"qwen2.5-32b","model_name":"Qwen2.5 32B Instruct","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"qwen2.5-72b","model_name":"Qwen2.5 72B Instruct","params_b":72,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"qwen3-8b","model_name":"Qwen3 8B","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":7.20000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"qwen3-14b","model_name":"Qwen3 14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":6.70000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"qwen3-32b","model_name":"Qwen3 32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"qwen3-30b-a3b","model_name":"Qwen3 30B-A3B (MoE)","params_b":30,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"qwen3-235b-a22b","model_name":"Qwen3 235B-A22B (MoE)","params_b":235,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"gemma3-4b","model_name":"Gemma 3 4B Instruct","params_b":4,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":7.79999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"gemma3-12b","model_name":"Gemma 3 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":3.20000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"gemma3-27b","model_name":"Gemma 3 27B Instruct","params_b":27,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"gemma4-e4b","model_name":"Gemma 4 E4B (Edge)","params_b":4.5,"q4_fits":true,"q4_quality":"ample","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"q4","headroom_at_recommended":6.4000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"gemma4-31b","model_name":"Gemma 4 31B","params_b":31,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"mistral-7b-v0.3","model_name":"Mistral 7B v0.3 Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"mistral-nemo-12b","model_name":"Mistral Nemo 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":3.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"mixtral-8x7b","model_name":"Mixtral 8x7B Instruct v0.1","params_b":46.7000000000000028421709430404007434844970703125,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"phi3.5-mini-3.8b","model_name":"Phi-3.5 mini 3.8B Instruct","params_b":3.79999999999999982236431605997495353221893310546875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":8.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"phi4-14b","model_name":"Phi-4 14B Instruct","params_b":14.699999999999999289457264239899814128875732421875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":6.9000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"deepseek-r1-distill-qwen-7b","model_name":"DeepSeek-R1-Distill-Qwen-7B","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"deepseek-r1-distill-qwen-14b","model_name":"DeepSeek-R1-Distill-Qwen-14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":7,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"deepseek-r1-distill-qwen-32b","model_name":"DeepSeek-R1-Distill-Qwen-32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"deepseek-r1-distill-llama-70b","model_name":"DeepSeek-R1-Distill-Llama-70B","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"olmo2-7b","model_name":"OLMo 2 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"olmo2-13b","model_name":"OLMo 2 13B Instruct","params_b":13,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":2.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"command-r-35b","model_name":"Command R 35B (08-2024)","params_b":35,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-ti-super-16gb","gpu_name":"NVIDIA GeForce RTX 4070 Ti Super","gpu_vram_gb":16,"model_id":"command-r-plus-104b","model_name":"Command R+ 104B","params_b":104,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"smollm2-1.7b","model_name":"SmolLM2 1.7B Instruct","params_b":1.6999999999999999555910790149937383830547332763671875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":16.39999999999999857891452847979962825775146484375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"llama3.2-1b","model_name":"Llama 3.2 1B Instruct","params_b":1,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":17.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"llama3.2-3b","model_name":"Llama 3.2 3B Instruct","params_b":3,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":13.699999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"llama3.1-8b","model_name":"Llama 3.1 8B Instruct","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":3.79999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"llama3.1-70b","model_name":"Llama 3.1 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"llama3.3-70b","model_name":"Llama 3.3 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"qwen2.5-7b","model_name":"Qwen2.5 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":5.79999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"qwen2.5-14b","model_name":"Qwen2.5 14B Instruct","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":5.20000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"qwen2.5-32b","model_name":"Qwen2.5 32B Instruct","params_b":32,"q4_fits":true,"q4_quality":"tight","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":1,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"qwen2.5-72b","model_name":"Qwen2.5 72B Instruct","params_b":72,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"qwen3-8b","model_name":"Qwen3 8B","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":3.600000000000000088817841970012523233890533447265625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"qwen3-14b","model_name":"Qwen3 14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":5.20000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"qwen3-32b","model_name":"Qwen3 32B","params_b":32,"q4_fits":true,"q4_quality":"tight","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":0,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"qwen3-30b-a3b","model_name":"Qwen3 30B-A3B (MoE)","params_b":30,"q4_fits":true,"q4_quality":"tight","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":1,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"qwen3-235b-a22b","model_name":"Qwen3 235B-A22B (MoE)","params_b":235,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"gemma3-4b","model_name":"Gemma 3 4B Instruct","params_b":4,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":11.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"gemma3-12b","model_name":"Gemma 3 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":7.20000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"gemma3-27b","model_name":"Gemma 3 27B Instruct","params_b":27,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"gemma4-e4b","model_name":"Gemma 4 E4B (Edge)","params_b":4.5,"q4_fits":true,"q4_quality":"ample","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"q4","headroom_at_recommended":10.4000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"gemma4-31b","model_name":"Gemma 4 31B","params_b":31,"q4_fits":true,"q4_quality":"tight","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":0,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"mistral-7b-v0.3","model_name":"Mistral 7B v0.3 Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":5.79999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"mistral-nemo-12b","model_name":"Mistral Nemo 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":7.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"mixtral-8x7b","model_name":"Mixtral 8x7B Instruct v0.1","params_b":46.7000000000000028421709430404007434844970703125,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"phi3.5-mini-3.8b","model_name":"Phi-3.5 mini 3.8B Instruct","params_b":3.79999999999999982236431605997495353221893310546875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":12.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"phi4-14b","model_name":"Phi-4 14B Instruct","params_b":14.699999999999999289457264239899814128875732421875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":4.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"deepseek-r1-distill-qwen-7b","model_name":"DeepSeek-R1-Distill-Qwen-7B","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":5.79999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"deepseek-r1-distill-qwen-14b","model_name":"DeepSeek-R1-Distill-Qwen-14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":5.20000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"deepseek-r1-distill-qwen-32b","model_name":"DeepSeek-R1-Distill-Qwen-32B","params_b":32,"q4_fits":true,"q4_quality":"tight","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":0,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"deepseek-r1-distill-llama-70b","model_name":"DeepSeek-R1-Distill-Llama-70B","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"olmo2-7b","model_name":"OLMo 2 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":5.79999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"olmo2-13b","model_name":"OLMo 2 13B Instruct","params_b":13,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":6.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"command-r-35b","model_name":"Command R 35B (08-2024)","params_b":35,"q4_fits":true,"q4_quality":"tight","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":1,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7900-xt-20gb","gpu_name":"AMD Radeon RX 7900 XT","gpu_vram_gb":20,"model_id":"command-r-plus-104b","model_name":"Command R+ 104B","params_b":104,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"smollm2-1.7b","model_name":"SmolLM2 1.7B Instruct","params_b":1.6999999999999999555910790149937383830547332763671875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":60.39999999999999857891452847979962825775146484375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"llama3.2-1b","model_name":"Llama 3.2 1B Instruct","params_b":1,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":61.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"llama3.2-3b","model_name":"Llama 3.2 3B Instruct","params_b":3,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":57.7000000000000028421709430404007434844970703125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"llama3.1-8b","model_name":"Llama 3.1 8B Instruct","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":47.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"llama3.1-70b","model_name":"Llama 3.1 70B Instruct","params_b":70,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":21,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"llama3.3-70b","model_name":"Llama 3.3 70B Instruct","params_b":70,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":21,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"qwen2.5-7b","model_name":"Qwen2.5 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":49.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"qwen2.5-14b","model_name":"Qwen2.5 14B Instruct","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":35.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"qwen2.5-32b","model_name":"Qwen2.5 32B Instruct","params_b":32,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":31,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"qwen2.5-72b","model_name":"Qwen2.5 72B Instruct","params_b":72,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":20,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"qwen3-8b","model_name":"Qwen3 8B","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":47.60000000000000142108547152020037174224853515625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"qwen3-14b","model_name":"Qwen3 14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":35.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"qwen3-32b","model_name":"Qwen3 32B","params_b":32,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":31,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"qwen3-30b-a3b","model_name":"Qwen3 30B-A3B (MoE)","params_b":30,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":32,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"qwen3-235b-a22b","model_name":"Qwen3 235B-A22B (MoE)","params_b":235,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"gemma3-4b","model_name":"Gemma 3 4B Instruct","params_b":4,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":55.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"gemma3-12b","model_name":"Gemma 3 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":39.60000000000000142108547152020037174224853515625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"gemma3-27b","model_name":"Gemma 3 27B Instruct","params_b":27,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":9.300000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"gemma4-e4b","model_name":"Gemma 4 E4B (Edge)","params_b":4.5,"q4_fits":true,"q4_quality":"ample","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"q4","headroom_at_recommended":54.39999999999999857891452847979962825775146484375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"gemma4-31b","model_name":"Gemma 4 31B","params_b":31,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":31.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"mistral-7b-v0.3","model_name":"Mistral 7B v0.3 Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":49.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"mistral-nemo-12b","model_name":"Mistral Nemo 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":39.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"mixtral-8x7b","model_name":"Mixtral 8x7B Instruct v0.1","params_b":46.7000000000000028421709430404007434844970703125,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":16.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"phi3.5-mini-3.8b","model_name":"Phi-3.5 mini 3.8B Instruct","params_b":3.79999999999999982236431605997495353221893310546875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":56.2000000000000028421709430404007434844970703125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"phi4-14b","model_name":"Phi-4 14B Instruct","params_b":14.699999999999999289457264239899814128875732421875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":34.2000000000000028421709430404007434844970703125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"deepseek-r1-distill-qwen-7b","model_name":"DeepSeek-R1-Distill-Qwen-7B","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":49.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"deepseek-r1-distill-qwen-14b","model_name":"DeepSeek-R1-Distill-Qwen-14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":35.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"deepseek-r1-distill-qwen-32b","model_name":"DeepSeek-R1-Distill-Qwen-32B","params_b":32,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":31,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"deepseek-r1-distill-llama-70b","model_name":"DeepSeek-R1-Distill-Llama-70B","params_b":70,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":21,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"olmo2-7b","model_name":"OLMo 2 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":49.7999999999999971578290569595992565155029296875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"olmo2-13b","model_name":"OLMo 2 13B Instruct","params_b":13,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":37.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"command-r-35b","model_name":"Command R 35B (08-2024)","params_b":35,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":28.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"apple-m4-pro-64gb","gpu_name":"Apple M4 Pro (20-core GPU, up to 64 GB)","gpu_vram_gb":64,"model_id":"command-r-plus-104b","model_name":"Command R+ 104B","params_b":104,"q4_fits":true,"q4_quality":"tight","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"smollm2-1.7b","model_name":"SmolLM2 1.7B Instruct","params_b":1.6999999999999999555910790149937383830547332763671875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":8.4000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"llama3.2-1b","model_name":"Llama 3.2 1B Instruct","params_b":1,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":9.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"llama3.2-3b","model_name":"Llama 3.2 3B Instruct","params_b":3,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":5.70000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"llama3.1-8b","model_name":"Llama 3.1 8B Instruct","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":3.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"llama3.1-70b","model_name":"Llama 3.1 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"llama3.3-70b","model_name":"Llama 3.3 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"qwen2.5-7b","model_name":"Qwen2.5 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":4.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"qwen2.5-14b","model_name":"Qwen2.5 14B Instruct","params_b":14,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"qwen2.5-32b","model_name":"Qwen2.5 32B Instruct","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"qwen2.5-72b","model_name":"Qwen2.5 72B Instruct","params_b":72,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"qwen3-8b","model_name":"Qwen3 8B","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":3.20000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"qwen3-14b","model_name":"Qwen3 14B","params_b":14,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":2.70000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"qwen3-32b","model_name":"Qwen3 32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"qwen3-30b-a3b","model_name":"Qwen3 30B-A3B (MoE)","params_b":30,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"qwen3-235b-a22b","model_name":"Qwen3 235B-A22B (MoE)","params_b":235,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"gemma3-4b","model_name":"Gemma 3 4B Instruct","params_b":4,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":3.79999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"gemma3-12b","model_name":"Gemma 3 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":4.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"gemma3-27b","model_name":"Gemma 3 27B Instruct","params_b":27,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"gemma4-e4b","model_name":"Gemma 4 E4B (Edge)","params_b":4.5,"q4_fits":true,"q4_quality":"good","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"q4","headroom_at_recommended":2.399999999999999911182158029987476766109466552734375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"gemma4-31b","model_name":"Gemma 4 31B","params_b":31,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"mistral-7b-v0.3","model_name":"Mistral 7B v0.3 Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":4.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"mistral-nemo-12b","model_name":"Mistral Nemo 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":4.9000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"mixtral-8x7b","model_name":"Mixtral 8x7B Instruct v0.1","params_b":46.7000000000000028421709430404007434844970703125,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"phi3.5-mini-3.8b","model_name":"Phi-3.5 mini 3.8B Instruct","params_b":3.79999999999999982236431605997495353221893310546875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":4.20000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"phi4-14b","model_name":"Phi-4 14B Instruct","params_b":14.699999999999999289457264239899814128875732421875,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":2.899999999999999911182158029987476766109466552734375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"deepseek-r1-distill-qwen-7b","model_name":"DeepSeek-R1-Distill-Qwen-7B","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":4.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"deepseek-r1-distill-qwen-14b","model_name":"DeepSeek-R1-Distill-Qwen-14B","params_b":14,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"deepseek-r1-distill-qwen-32b","model_name":"DeepSeek-R1-Distill-Qwen-32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"deepseek-r1-distill-llama-70b","model_name":"DeepSeek-R1-Distill-Llama-70B","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"olmo2-7b","model_name":"OLMo 2 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":4.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"olmo2-13b","model_name":"OLMo 2 13B Instruct","params_b":13,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3.600000000000000088817841970012523233890533447265625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"command-r-35b","model_name":"Command R 35B (08-2024)","params_b":35,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-super-12gb","gpu_name":"NVIDIA GeForce RTX 4070 Super","gpu_vram_gb":12,"model_id":"command-r-plus-104b","model_name":"Command R+ 104B","params_b":104,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"smollm2-1.7b","model_name":"SmolLM2 1.7B Instruct","params_b":1.6999999999999999555910790149937383830547332763671875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":12.4000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"llama3.2-1b","model_name":"Llama 3.2 1B Instruct","params_b":1,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":13.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"llama3.2-3b","model_name":"Llama 3.2 3B Instruct","params_b":3,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":9.699999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"llama3.1-8b","model_name":"Llama 3.1 8B Instruct","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":7.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"llama3.1-70b","model_name":"Llama 3.1 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"llama3.3-70b","model_name":"Llama 3.3 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"qwen2.5-7b","model_name":"Qwen2.5 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"qwen2.5-14b","model_name":"Qwen2.5 14B Instruct","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":7.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"qwen2.5-32b","model_name":"Qwen2.5 32B Instruct","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"qwen2.5-72b","model_name":"Qwen2.5 72B Instruct","params_b":72,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"qwen3-8b","model_name":"Qwen3 8B","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":7.20000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"qwen3-14b","model_name":"Qwen3 14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":6.70000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"qwen3-32b","model_name":"Qwen3 32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"qwen3-30b-a3b","model_name":"Qwen3 30B-A3B (MoE)","params_b":30,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"qwen3-235b-a22b","model_name":"Qwen3 235B-A22B (MoE)","params_b":235,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"gemma3-4b","model_name":"Gemma 3 4B Instruct","params_b":4,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":7.79999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"gemma3-12b","model_name":"Gemma 3 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":3.20000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"gemma3-27b","model_name":"Gemma 3 27B Instruct","params_b":27,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"gemma4-e4b","model_name":"Gemma 4 E4B (Edge)","params_b":4.5,"q4_fits":true,"q4_quality":"ample","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"q4","headroom_at_recommended":6.4000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"gemma4-31b","model_name":"Gemma 4 31B","params_b":31,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"mistral-7b-v0.3","model_name":"Mistral 7B v0.3 Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"mistral-nemo-12b","model_name":"Mistral Nemo 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":3.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"mixtral-8x7b","model_name":"Mixtral 8x7B Instruct v0.1","params_b":46.7000000000000028421709430404007434844970703125,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"phi3.5-mini-3.8b","model_name":"Phi-3.5 mini 3.8B Instruct","params_b":3.79999999999999982236431605997495353221893310546875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":8.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"phi4-14b","model_name":"Phi-4 14B Instruct","params_b":14.699999999999999289457264239899814128875732421875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":6.9000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"deepseek-r1-distill-qwen-7b","model_name":"DeepSeek-R1-Distill-Qwen-7B","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"deepseek-r1-distill-qwen-14b","model_name":"DeepSeek-R1-Distill-Qwen-14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":7,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"deepseek-r1-distill-qwen-32b","model_name":"DeepSeek-R1-Distill-Qwen-32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"deepseek-r1-distill-llama-70b","model_name":"DeepSeek-R1-Distill-Llama-70B","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"olmo2-7b","model_name":"OLMo 2 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"olmo2-13b","model_name":"OLMo 2 13B Instruct","params_b":13,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":2.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"command-r-35b","model_name":"Command R 35B (08-2024)","params_b":35,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7800-xt-16gb","gpu_name":"AMD Radeon RX 7800 XT","gpu_vram_gb":16,"model_id":"command-r-plus-104b","model_name":"Command R+ 104B","params_b":104,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"smollm2-1.7b","model_name":"SmolLM2 1.7B Instruct","params_b":1.6999999999999999555910790149937383830547332763671875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":12.4000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"llama3.2-1b","model_name":"Llama 3.2 1B Instruct","params_b":1,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":13.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"llama3.2-3b","model_name":"Llama 3.2 3B Instruct","params_b":3,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":9.699999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"llama3.1-8b","model_name":"Llama 3.1 8B Instruct","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":7.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"llama3.1-70b","model_name":"Llama 3.1 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"llama3.3-70b","model_name":"Llama 3.3 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"qwen2.5-7b","model_name":"Qwen2.5 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"qwen2.5-14b","model_name":"Qwen2.5 14B Instruct","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":7.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"qwen2.5-32b","model_name":"Qwen2.5 32B Instruct","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"qwen2.5-72b","model_name":"Qwen2.5 72B Instruct","params_b":72,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"qwen3-8b","model_name":"Qwen3 8B","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":7.20000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"qwen3-14b","model_name":"Qwen3 14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":6.70000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"qwen3-32b","model_name":"Qwen3 32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"qwen3-30b-a3b","model_name":"Qwen3 30B-A3B (MoE)","params_b":30,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"qwen3-235b-a22b","model_name":"Qwen3 235B-A22B (MoE)","params_b":235,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"gemma3-4b","model_name":"Gemma 3 4B Instruct","params_b":4,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":7.79999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"gemma3-12b","model_name":"Gemma 3 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":3.20000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"gemma3-27b","model_name":"Gemma 3 27B Instruct","params_b":27,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"gemma4-e4b","model_name":"Gemma 4 E4B (Edge)","params_b":4.5,"q4_fits":true,"q4_quality":"ample","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"q4","headroom_at_recommended":6.4000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"gemma4-31b","model_name":"Gemma 4 31B","params_b":31,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"mistral-7b-v0.3","model_name":"Mistral 7B v0.3 Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"mistral-nemo-12b","model_name":"Mistral Nemo 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":3.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"mixtral-8x7b","model_name":"Mixtral 8x7B Instruct v0.1","params_b":46.7000000000000028421709430404007434844970703125,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"phi3.5-mini-3.8b","model_name":"Phi-3.5 mini 3.8B Instruct","params_b":3.79999999999999982236431605997495353221893310546875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":8.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"phi4-14b","model_name":"Phi-4 14B Instruct","params_b":14.699999999999999289457264239899814128875732421875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":6.9000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"deepseek-r1-distill-qwen-7b","model_name":"DeepSeek-R1-Distill-Qwen-7B","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"deepseek-r1-distill-qwen-14b","model_name":"DeepSeek-R1-Distill-Qwen-14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":7,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"deepseek-r1-distill-qwen-32b","model_name":"DeepSeek-R1-Distill-Qwen-32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"deepseek-r1-distill-llama-70b","model_name":"DeepSeek-R1-Distill-Llama-70B","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"olmo2-7b","model_name":"OLMo 2 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"olmo2-13b","model_name":"OLMo 2 13B Instruct","params_b":13,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":2.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"command-r-35b","model_name":"Command R 35B (08-2024)","params_b":35,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"arc-a770-16gb","gpu_name":"Intel Arc A770 16 GB","gpu_vram_gb":16,"model_id":"command-r-plus-104b","model_name":"Command R+ 104B","params_b":104,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"smollm2-1.7b","model_name":"SmolLM2 1.7B Instruct","params_b":1.6999999999999999555910790149937383830547332763671875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":8.4000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"llama3.2-1b","model_name":"Llama 3.2 1B Instruct","params_b":1,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":9.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"llama3.2-3b","model_name":"Llama 3.2 3B Instruct","params_b":3,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":5.70000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"llama3.1-8b","model_name":"Llama 3.1 8B Instruct","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":3.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"llama3.1-70b","model_name":"Llama 3.1 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"llama3.3-70b","model_name":"Llama 3.3 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"qwen2.5-7b","model_name":"Qwen2.5 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":4.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"qwen2.5-14b","model_name":"Qwen2.5 14B Instruct","params_b":14,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"qwen2.5-32b","model_name":"Qwen2.5 32B Instruct","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"qwen2.5-72b","model_name":"Qwen2.5 72B Instruct","params_b":72,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"qwen3-8b","model_name":"Qwen3 8B","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":3.20000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"qwen3-14b","model_name":"Qwen3 14B","params_b":14,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":2.70000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"qwen3-32b","model_name":"Qwen3 32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"qwen3-30b-a3b","model_name":"Qwen3 30B-A3B (MoE)","params_b":30,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"qwen3-235b-a22b","model_name":"Qwen3 235B-A22B (MoE)","params_b":235,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"gemma3-4b","model_name":"Gemma 3 4B Instruct","params_b":4,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":3.79999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"gemma3-12b","model_name":"Gemma 3 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":4.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"gemma3-27b","model_name":"Gemma 3 27B Instruct","params_b":27,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"gemma4-e4b","model_name":"Gemma 4 E4B (Edge)","params_b":4.5,"q4_fits":true,"q4_quality":"good","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"q4","headroom_at_recommended":2.399999999999999911182158029987476766109466552734375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"gemma4-31b","model_name":"Gemma 4 31B","params_b":31,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"mistral-7b-v0.3","model_name":"Mistral 7B v0.3 Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":4.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"mistral-nemo-12b","model_name":"Mistral Nemo 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":4.9000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"mixtral-8x7b","model_name":"Mixtral 8x7B Instruct v0.1","params_b":46.7000000000000028421709430404007434844970703125,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"phi3.5-mini-3.8b","model_name":"Phi-3.5 mini 3.8B Instruct","params_b":3.79999999999999982236431605997495353221893310546875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":4.20000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"phi4-14b","model_name":"Phi-4 14B Instruct","params_b":14.699999999999999289457264239899814128875732421875,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":2.899999999999999911182158029987476766109466552734375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"deepseek-r1-distill-qwen-7b","model_name":"DeepSeek-R1-Distill-Qwen-7B","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":4.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"deepseek-r1-distill-qwen-14b","model_name":"DeepSeek-R1-Distill-Qwen-14B","params_b":14,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"deepseek-r1-distill-qwen-32b","model_name":"DeepSeek-R1-Distill-Qwen-32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"deepseek-r1-distill-llama-70b","model_name":"DeepSeek-R1-Distill-Llama-70B","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"olmo2-7b","model_name":"OLMo 2 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":4.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"olmo2-13b","model_name":"OLMo 2 13B Instruct","params_b":13,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3.600000000000000088817841970012523233890533447265625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"command-r-35b","model_name":"Command R 35B (08-2024)","params_b":35,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7700-xt-12gb","gpu_name":"AMD Radeon RX 7700 XT","gpu_vram_gb":12,"model_id":"command-r-plus-104b","model_name":"Command R+ 104B","params_b":104,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"smollm2-1.7b","model_name":"SmolLM2 1.7B Instruct","params_b":1.6999999999999999555910790149937383830547332763671875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":8.4000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"llama3.2-1b","model_name":"Llama 3.2 1B Instruct","params_b":1,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":9.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"llama3.2-3b","model_name":"Llama 3.2 3B Instruct","params_b":3,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":5.70000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"llama3.1-8b","model_name":"Llama 3.1 8B Instruct","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":3.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"llama3.1-70b","model_name":"Llama 3.1 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"llama3.3-70b","model_name":"Llama 3.3 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"qwen2.5-7b","model_name":"Qwen2.5 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":4.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"qwen2.5-14b","model_name":"Qwen2.5 14B Instruct","params_b":14,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"qwen2.5-32b","model_name":"Qwen2.5 32B Instruct","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"qwen2.5-72b","model_name":"Qwen2.5 72B Instruct","params_b":72,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"qwen3-8b","model_name":"Qwen3 8B","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":3.20000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"qwen3-14b","model_name":"Qwen3 14B","params_b":14,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":2.70000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"qwen3-32b","model_name":"Qwen3 32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"qwen3-30b-a3b","model_name":"Qwen3 30B-A3B (MoE)","params_b":30,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"qwen3-235b-a22b","model_name":"Qwen3 235B-A22B (MoE)","params_b":235,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"gemma3-4b","model_name":"Gemma 3 4B Instruct","params_b":4,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":3.79999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"gemma3-12b","model_name":"Gemma 3 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":4.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"gemma3-27b","model_name":"Gemma 3 27B Instruct","params_b":27,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"gemma4-e4b","model_name":"Gemma 4 E4B (Edge)","params_b":4.5,"q4_fits":true,"q4_quality":"good","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"q4","headroom_at_recommended":2.399999999999999911182158029987476766109466552734375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"gemma4-31b","model_name":"Gemma 4 31B","params_b":31,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"mistral-7b-v0.3","model_name":"Mistral 7B v0.3 Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":4.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"mistral-nemo-12b","model_name":"Mistral Nemo 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":4.9000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"mixtral-8x7b","model_name":"Mixtral 8x7B Instruct v0.1","params_b":46.7000000000000028421709430404007434844970703125,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"phi3.5-mini-3.8b","model_name":"Phi-3.5 mini 3.8B Instruct","params_b":3.79999999999999982236431605997495353221893310546875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":4.20000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"phi4-14b","model_name":"Phi-4 14B Instruct","params_b":14.699999999999999289457264239899814128875732421875,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":2.899999999999999911182158029987476766109466552734375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"deepseek-r1-distill-qwen-7b","model_name":"DeepSeek-R1-Distill-Qwen-7B","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":4.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"deepseek-r1-distill-qwen-14b","model_name":"DeepSeek-R1-Distill-Qwen-14B","params_b":14,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"deepseek-r1-distill-qwen-32b","model_name":"DeepSeek-R1-Distill-Qwen-32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"deepseek-r1-distill-llama-70b","model_name":"DeepSeek-R1-Distill-Llama-70B","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"olmo2-7b","model_name":"OLMo 2 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":4.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"olmo2-13b","model_name":"OLMo 2 13B Instruct","params_b":13,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3.600000000000000088817841970012523233890533447265625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"command-r-35b","model_name":"Command R 35B (08-2024)","params_b":35,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4070-12gb","gpu_name":"NVIDIA GeForce RTX 4070","gpu_vram_gb":12,"model_id":"command-r-plus-104b","model_name":"Command R+ 104B","params_b":104,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"smollm2-1.7b","model_name":"SmolLM2 1.7B Instruct","params_b":1.6999999999999999555910790149937383830547332763671875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":12.4000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"llama3.2-1b","model_name":"Llama 3.2 1B Instruct","params_b":1,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":13.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"llama3.2-3b","model_name":"Llama 3.2 3B Instruct","params_b":3,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":9.699999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"llama3.1-8b","model_name":"Llama 3.1 8B Instruct","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":7.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"llama3.1-70b","model_name":"Llama 3.1 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"llama3.3-70b","model_name":"Llama 3.3 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"qwen2.5-7b","model_name":"Qwen2.5 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"qwen2.5-14b","model_name":"Qwen2.5 14B Instruct","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":7.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"qwen2.5-32b","model_name":"Qwen2.5 32B Instruct","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"qwen2.5-72b","model_name":"Qwen2.5 72B Instruct","params_b":72,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"qwen3-8b","model_name":"Qwen3 8B","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":7.20000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"qwen3-14b","model_name":"Qwen3 14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":6.70000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"qwen3-32b","model_name":"Qwen3 32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"qwen3-30b-a3b","model_name":"Qwen3 30B-A3B (MoE)","params_b":30,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"qwen3-235b-a22b","model_name":"Qwen3 235B-A22B (MoE)","params_b":235,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"gemma3-4b","model_name":"Gemma 3 4B Instruct","params_b":4,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":7.79999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"gemma3-12b","model_name":"Gemma 3 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":3.20000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"gemma3-27b","model_name":"Gemma 3 27B Instruct","params_b":27,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"gemma4-e4b","model_name":"Gemma 4 E4B (Edge)","params_b":4.5,"q4_fits":true,"q4_quality":"ample","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"q4","headroom_at_recommended":6.4000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"gemma4-31b","model_name":"Gemma 4 31B","params_b":31,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"mistral-7b-v0.3","model_name":"Mistral 7B v0.3 Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"mistral-nemo-12b","model_name":"Mistral Nemo 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":3.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"mixtral-8x7b","model_name":"Mixtral 8x7B Instruct v0.1","params_b":46.7000000000000028421709430404007434844970703125,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"phi3.5-mini-3.8b","model_name":"Phi-3.5 mini 3.8B Instruct","params_b":3.79999999999999982236431605997495353221893310546875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":8.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"phi4-14b","model_name":"Phi-4 14B Instruct","params_b":14.699999999999999289457264239899814128875732421875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":6.9000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"deepseek-r1-distill-qwen-7b","model_name":"DeepSeek-R1-Distill-Qwen-7B","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"deepseek-r1-distill-qwen-14b","model_name":"DeepSeek-R1-Distill-Qwen-14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":7,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"deepseek-r1-distill-qwen-32b","model_name":"DeepSeek-R1-Distill-Qwen-32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"deepseek-r1-distill-llama-70b","model_name":"DeepSeek-R1-Distill-Llama-70B","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"olmo2-7b","model_name":"OLMo 2 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"olmo2-13b","model_name":"OLMo 2 13B Instruct","params_b":13,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":2.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"command-r-35b","model_name":"Command R 35B (08-2024)","params_b":35,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-16gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 16 GB","gpu_vram_gb":16,"model_id":"command-r-plus-104b","model_name":"Command R+ 104B","params_b":104,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"smollm2-1.7b","model_name":"SmolLM2 1.7B Instruct","params_b":1.6999999999999999555910790149937383830547332763671875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":8.4000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"llama3.2-1b","model_name":"Llama 3.2 1B Instruct","params_b":1,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":9.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"llama3.2-3b","model_name":"Llama 3.2 3B Instruct","params_b":3,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":5.70000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"llama3.1-8b","model_name":"Llama 3.1 8B Instruct","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":3.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"llama3.1-70b","model_name":"Llama 3.1 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"llama3.3-70b","model_name":"Llama 3.3 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"qwen2.5-7b","model_name":"Qwen2.5 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":4.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"qwen2.5-14b","model_name":"Qwen2.5 14B Instruct","params_b":14,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"qwen2.5-32b","model_name":"Qwen2.5 32B Instruct","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"qwen2.5-72b","model_name":"Qwen2.5 72B Instruct","params_b":72,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"qwen3-8b","model_name":"Qwen3 8B","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":3.20000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"qwen3-14b","model_name":"Qwen3 14B","params_b":14,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":2.70000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"qwen3-32b","model_name":"Qwen3 32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"qwen3-30b-a3b","model_name":"Qwen3 30B-A3B (MoE)","params_b":30,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"qwen3-235b-a22b","model_name":"Qwen3 235B-A22B (MoE)","params_b":235,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"gemma3-4b","model_name":"Gemma 3 4B Instruct","params_b":4,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":3.79999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"gemma3-12b","model_name":"Gemma 3 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":4.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"gemma3-27b","model_name":"Gemma 3 27B Instruct","params_b":27,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"gemma4-e4b","model_name":"Gemma 4 E4B (Edge)","params_b":4.5,"q4_fits":true,"q4_quality":"good","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"q4","headroom_at_recommended":2.399999999999999911182158029987476766109466552734375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"gemma4-31b","model_name":"Gemma 4 31B","params_b":31,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"mistral-7b-v0.3","model_name":"Mistral 7B v0.3 Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":4.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"mistral-nemo-12b","model_name":"Mistral Nemo 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":4.9000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"mixtral-8x7b","model_name":"Mixtral 8x7B Instruct v0.1","params_b":46.7000000000000028421709430404007434844970703125,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"phi3.5-mini-3.8b","model_name":"Phi-3.5 mini 3.8B Instruct","params_b":3.79999999999999982236431605997495353221893310546875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":4.20000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"phi4-14b","model_name":"Phi-4 14B Instruct","params_b":14.699999999999999289457264239899814128875732421875,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":2.899999999999999911182158029987476766109466552734375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"deepseek-r1-distill-qwen-7b","model_name":"DeepSeek-R1-Distill-Qwen-7B","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":4.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"deepseek-r1-distill-qwen-14b","model_name":"DeepSeek-R1-Distill-Qwen-14B","params_b":14,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"deepseek-r1-distill-qwen-32b","model_name":"DeepSeek-R1-Distill-Qwen-32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"deepseek-r1-distill-llama-70b","model_name":"DeepSeek-R1-Distill-Llama-70B","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"olmo2-7b","model_name":"OLMo 2 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":4.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"olmo2-13b","model_name":"OLMo 2 13B Instruct","params_b":13,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3.600000000000000088817841970012523233890533447265625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"command-r-35b","model_name":"Command R 35B (08-2024)","params_b":35,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3080-12gb","gpu_name":"NVIDIA GeForce RTX 3080 12 GB","gpu_vram_gb":12,"model_id":"command-r-plus-104b","model_name":"Command R+ 104B","params_b":104,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"smollm2-1.7b","model_name":"SmolLM2 1.7B Instruct","params_b":1.6999999999999999555910790149937383830547332763671875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":12.4000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"llama3.2-1b","model_name":"Llama 3.2 1B Instruct","params_b":1,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":13.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"llama3.2-3b","model_name":"Llama 3.2 3B Instruct","params_b":3,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":9.699999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"llama3.1-8b","model_name":"Llama 3.1 8B Instruct","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":7.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"llama3.1-70b","model_name":"Llama 3.1 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"llama3.3-70b","model_name":"Llama 3.3 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"qwen2.5-7b","model_name":"Qwen2.5 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"qwen2.5-14b","model_name":"Qwen2.5 14B Instruct","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":7.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"qwen2.5-32b","model_name":"Qwen2.5 32B Instruct","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"qwen2.5-72b","model_name":"Qwen2.5 72B Instruct","params_b":72,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"qwen3-8b","model_name":"Qwen3 8B","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":7.20000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"qwen3-14b","model_name":"Qwen3 14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":6.70000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"qwen3-32b","model_name":"Qwen3 32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"qwen3-30b-a3b","model_name":"Qwen3 30B-A3B (MoE)","params_b":30,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"qwen3-235b-a22b","model_name":"Qwen3 235B-A22B (MoE)","params_b":235,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"gemma3-4b","model_name":"Gemma 3 4B Instruct","params_b":4,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":7.79999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"gemma3-12b","model_name":"Gemma 3 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":3.20000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"gemma3-27b","model_name":"Gemma 3 27B Instruct","params_b":27,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"gemma4-e4b","model_name":"Gemma 4 E4B (Edge)","params_b":4.5,"q4_fits":true,"q4_quality":"ample","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"q4","headroom_at_recommended":6.4000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"gemma4-31b","model_name":"Gemma 4 31B","params_b":31,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"mistral-7b-v0.3","model_name":"Mistral 7B v0.3 Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"mistral-nemo-12b","model_name":"Mistral Nemo 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":3.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"mixtral-8x7b","model_name":"Mixtral 8x7B Instruct v0.1","params_b":46.7000000000000028421709430404007434844970703125,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"phi3.5-mini-3.8b","model_name":"Phi-3.5 mini 3.8B Instruct","params_b":3.79999999999999982236431605997495353221893310546875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":8.199999999999999289457264239899814128875732421875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"phi4-14b","model_name":"Phi-4 14B Instruct","params_b":14.699999999999999289457264239899814128875732421875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":6.9000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"deepseek-r1-distill-qwen-7b","model_name":"DeepSeek-R1-Distill-Qwen-7B","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"deepseek-r1-distill-qwen-14b","model_name":"DeepSeek-R1-Distill-Qwen-14B","params_b":14,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":7,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"deepseek-r1-distill-qwen-32b","model_name":"DeepSeek-R1-Distill-Qwen-32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"deepseek-r1-distill-llama-70b","model_name":"DeepSeek-R1-Distill-Llama-70B","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"olmo2-7b","model_name":"OLMo 2 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":8.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"olmo2-13b","model_name":"OLMo 2 13B Instruct","params_b":13,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":2.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"command-r-35b","model_name":"Command R 35B (08-2024)","params_b":35,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-6900-xt-16gb","gpu_name":"AMD Radeon RX 6900 XT","gpu_vram_gb":16,"model_id":"command-r-plus-104b","model_name":"Command R+ 104B","params_b":104,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"smollm2-1.7b","model_name":"SmolLM2 1.7B Instruct","params_b":1.6999999999999999555910790149937383830547332763671875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":8.4000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"llama3.2-1b","model_name":"Llama 3.2 1B Instruct","params_b":1,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":9.800000000000000710542735760100185871124267578125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"llama3.2-3b","model_name":"Llama 3.2 3B Instruct","params_b":3,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":5.70000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"llama3.1-8b","model_name":"Llama 3.1 8B Instruct","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":3.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"llama3.1-70b","model_name":"Llama 3.1 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"llama3.3-70b","model_name":"Llama 3.3 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"qwen2.5-7b","model_name":"Qwen2.5 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":4.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"qwen2.5-14b","model_name":"Qwen2.5 14B Instruct","params_b":14,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"qwen2.5-32b","model_name":"Qwen2.5 32B Instruct","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"qwen2.5-72b","model_name":"Qwen2.5 72B Instruct","params_b":72,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"qwen3-8b","model_name":"Qwen3 8B","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"good","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":3.20000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"qwen3-14b","model_name":"Qwen3 14B","params_b":14,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":2.70000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"qwen3-32b","model_name":"Qwen3 32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"qwen3-30b-a3b","model_name":"Qwen3 30B-A3B (MoE)","params_b":30,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"qwen3-235b-a22b","model_name":"Qwen3 235B-A22B (MoE)","params_b":235,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"gemma3-4b","model_name":"Gemma 3 4B Instruct","params_b":4,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":3.79999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"gemma3-12b","model_name":"Gemma 3 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":4.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"gemma3-27b","model_name":"Gemma 3 27B Instruct","params_b":27,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"gemma4-e4b","model_name":"Gemma 4 E4B (Edge)","params_b":4.5,"q4_fits":true,"q4_quality":"good","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"q4","headroom_at_recommended":2.399999999999999911182158029987476766109466552734375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"gemma4-31b","model_name":"Gemma 4 31B","params_b":31,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"mistral-7b-v0.3","model_name":"Mistral 7B v0.3 Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":4.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"mistral-nemo-12b","model_name":"Mistral Nemo 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"ample","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":4.9000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"mixtral-8x7b","model_name":"Mixtral 8x7B Instruct v0.1","params_b":46.7000000000000028421709430404007434844970703125,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"phi3.5-mini-3.8b","model_name":"Phi-3.5 mini 3.8B Instruct","params_b":3.79999999999999982236431605997495353221893310546875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":4.20000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"phi4-14b","model_name":"Phi-4 14B Instruct","params_b":14.699999999999999289457264239899814128875732421875,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":2.899999999999999911182158029987476766109466552734375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"deepseek-r1-distill-qwen-7b","model_name":"DeepSeek-R1-Distill-Qwen-7B","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":4.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"deepseek-r1-distill-qwen-14b","model_name":"DeepSeek-R1-Distill-Qwen-14B","params_b":14,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"deepseek-r1-distill-qwen-32b","model_name":"DeepSeek-R1-Distill-Qwen-32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"deepseek-r1-distill-llama-70b","model_name":"DeepSeek-R1-Distill-Llama-70B","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"olmo2-7b","model_name":"OLMo 2 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":4.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"olmo2-13b","model_name":"OLMo 2 13B Instruct","params_b":13,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3.600000000000000088817841970012523233890533447265625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"command-r-35b","model_name":"Command R 35B (08-2024)","params_b":35,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"arc-b580-12gb","gpu_name":"Intel Arc B580 12 GB","gpu_vram_gb":12,"model_id":"command-r-plus-104b","model_name":"Command R+ 104B","params_b":104,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"smollm2-1.7b","model_name":"SmolLM2 1.7B Instruct","params_b":1.6999999999999999555910790149937383830547332763671875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":4.4000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"llama3.2-1b","model_name":"Llama 3.2 1B Instruct","params_b":1,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":5.79999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"llama3.2-3b","model_name":"Llama 3.2 3B Instruct","params_b":3,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":1.6999999999999999555910790149937383830547332763671875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"llama3.1-8b","model_name":"Llama 3.1 8B Instruct","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"llama3.1-70b","model_name":"Llama 3.1 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"llama3.3-70b","model_name":"Llama 3.3 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"qwen2.5-7b","model_name":"Qwen2.5 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"qwen2.5-14b","model_name":"Qwen2.5 14B Instruct","params_b":14,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"qwen2.5-32b","model_name":"Qwen2.5 32B Instruct","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"qwen2.5-72b","model_name":"Qwen2.5 72B Instruct","params_b":72,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"qwen3-8b","model_name":"Qwen3 8B","params_b":8,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":2.79999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"qwen3-14b","model_name":"Qwen3 14B","params_b":14,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"qwen3-32b","model_name":"Qwen3 32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"qwen3-30b-a3b","model_name":"Qwen3 30B-A3B (MoE)","params_b":30,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"qwen3-235b-a22b","model_name":"Qwen3 235B-A22B (MoE)","params_b":235,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"gemma3-4b","model_name":"Gemma 3 4B Instruct","params_b":4,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":3.70000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"gemma3-12b","model_name":"Gemma 3 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"tight","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":0.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"gemma3-27b","model_name":"Gemma 3 27B Instruct","params_b":27,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"gemma4-e4b","model_name":"Gemma 4 E4B (Edge)","params_b":4.5,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"gemma4-31b","model_name":"Gemma 4 31B","params_b":31,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"mistral-7b-v0.3","model_name":"Mistral 7B v0.3 Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"mistral-nemo-12b","model_name":"Mistral Nemo 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"tight","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":0.90000000000000002220446049250313080847263336181640625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"mixtral-8x7b","model_name":"Mixtral 8x7B Instruct v0.1","params_b":46.7000000000000028421709430404007434844970703125,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"phi3.5-mini-3.8b","model_name":"Phi-3.5 mini 3.8B Instruct","params_b":3.79999999999999982236431605997495353221893310546875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":3.899999999999999911182158029987476766109466552734375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"phi4-14b","model_name":"Phi-4 14B Instruct","params_b":14.699999999999999289457264239899814128875732421875,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"deepseek-r1-distill-qwen-7b","model_name":"DeepSeek-R1-Distill-Qwen-7B","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3.29999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"deepseek-r1-distill-qwen-14b","model_name":"DeepSeek-R1-Distill-Qwen-14B","params_b":14,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"deepseek-r1-distill-qwen-32b","model_name":"DeepSeek-R1-Distill-Qwen-32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"deepseek-r1-distill-llama-70b","model_name":"DeepSeek-R1-Distill-Llama-70B","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"olmo2-7b","model_name":"OLMo 2 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"olmo2-13b","model_name":"OLMo 2 13B Instruct","params_b":13,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"command-r-35b","model_name":"Command R 35B (08-2024)","params_b":35,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-ti-8gb","gpu_name":"NVIDIA GeForce RTX 4060 Ti 8 GB","gpu_vram_gb":8,"model_id":"command-r-plus-104b","model_name":"Command R+ 104B","params_b":104,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"smollm2-1.7b","model_name":"SmolLM2 1.7B Instruct","params_b":1.6999999999999999555910790149937383830547332763671875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":4.4000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"llama3.2-1b","model_name":"Llama 3.2 1B Instruct","params_b":1,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":5.79999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"llama3.2-3b","model_name":"Llama 3.2 3B Instruct","params_b":3,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":1.6999999999999999555910790149937383830547332763671875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"llama3.1-8b","model_name":"Llama 3.1 8B Instruct","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"llama3.1-70b","model_name":"Llama 3.1 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"llama3.3-70b","model_name":"Llama 3.3 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"qwen2.5-7b","model_name":"Qwen2.5 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"qwen2.5-14b","model_name":"Qwen2.5 14B Instruct","params_b":14,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"qwen2.5-32b","model_name":"Qwen2.5 32B Instruct","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"qwen2.5-72b","model_name":"Qwen2.5 72B Instruct","params_b":72,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"qwen3-8b","model_name":"Qwen3 8B","params_b":8,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":2.79999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"qwen3-14b","model_name":"Qwen3 14B","params_b":14,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"qwen3-32b","model_name":"Qwen3 32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"qwen3-30b-a3b","model_name":"Qwen3 30B-A3B (MoE)","params_b":30,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"qwen3-235b-a22b","model_name":"Qwen3 235B-A22B (MoE)","params_b":235,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"gemma3-4b","model_name":"Gemma 3 4B Instruct","params_b":4,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":3.70000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"gemma3-12b","model_name":"Gemma 3 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"tight","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":0.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"gemma3-27b","model_name":"Gemma 3 27B Instruct","params_b":27,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"gemma4-e4b","model_name":"Gemma 4 E4B (Edge)","params_b":4.5,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"gemma4-31b","model_name":"Gemma 4 31B","params_b":31,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"mistral-7b-v0.3","model_name":"Mistral 7B v0.3 Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"mistral-nemo-12b","model_name":"Mistral Nemo 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"tight","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":0.90000000000000002220446049250313080847263336181640625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"mixtral-8x7b","model_name":"Mixtral 8x7B Instruct v0.1","params_b":46.7000000000000028421709430404007434844970703125,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"phi3.5-mini-3.8b","model_name":"Phi-3.5 mini 3.8B Instruct","params_b":3.79999999999999982236431605997495353221893310546875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":3.899999999999999911182158029987476766109466552734375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"phi4-14b","model_name":"Phi-4 14B Instruct","params_b":14.699999999999999289457264239899814128875732421875,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"deepseek-r1-distill-qwen-7b","model_name":"DeepSeek-R1-Distill-Qwen-7B","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3.29999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"deepseek-r1-distill-qwen-14b","model_name":"DeepSeek-R1-Distill-Qwen-14B","params_b":14,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"deepseek-r1-distill-qwen-32b","model_name":"DeepSeek-R1-Distill-Qwen-32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"deepseek-r1-distill-llama-70b","model_name":"DeepSeek-R1-Distill-Llama-70B","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"olmo2-7b","model_name":"OLMo 2 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"olmo2-13b","model_name":"OLMo 2 13B Instruct","params_b":13,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"command-r-35b","model_name":"Command R 35B (08-2024)","params_b":35,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-4060-8gb","gpu_name":"NVIDIA GeForce RTX 4060","gpu_vram_gb":8,"model_id":"command-r-plus-104b","model_name":"Command R+ 104B","params_b":104,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"smollm2-1.7b","model_name":"SmolLM2 1.7B Instruct","params_b":1.6999999999999999555910790149937383830547332763671875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":4.4000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"llama3.2-1b","model_name":"Llama 3.2 1B Instruct","params_b":1,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":5.79999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"llama3.2-3b","model_name":"Llama 3.2 3B Instruct","params_b":3,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":1.6999999999999999555910790149937383830547332763671875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"llama3.1-8b","model_name":"Llama 3.1 8B Instruct","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"llama3.1-70b","model_name":"Llama 3.1 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"llama3.3-70b","model_name":"Llama 3.3 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"qwen2.5-7b","model_name":"Qwen2.5 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"qwen2.5-14b","model_name":"Qwen2.5 14B Instruct","params_b":14,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"qwen2.5-32b","model_name":"Qwen2.5 32B Instruct","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"qwen2.5-72b","model_name":"Qwen2.5 72B Instruct","params_b":72,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"qwen3-8b","model_name":"Qwen3 8B","params_b":8,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":2.79999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"qwen3-14b","model_name":"Qwen3 14B","params_b":14,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"qwen3-32b","model_name":"Qwen3 32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"qwen3-30b-a3b","model_name":"Qwen3 30B-A3B (MoE)","params_b":30,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"qwen3-235b-a22b","model_name":"Qwen3 235B-A22B (MoE)","params_b":235,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"gemma3-4b","model_name":"Gemma 3 4B Instruct","params_b":4,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":3.70000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"gemma3-12b","model_name":"Gemma 3 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"tight","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":0.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"gemma3-27b","model_name":"Gemma 3 27B Instruct","params_b":27,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"gemma4-e4b","model_name":"Gemma 4 E4B (Edge)","params_b":4.5,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"gemma4-31b","model_name":"Gemma 4 31B","params_b":31,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"mistral-7b-v0.3","model_name":"Mistral 7B v0.3 Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"mistral-nemo-12b","model_name":"Mistral Nemo 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"tight","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":0.90000000000000002220446049250313080847263336181640625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"mixtral-8x7b","model_name":"Mixtral 8x7B Instruct v0.1","params_b":46.7000000000000028421709430404007434844970703125,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"phi3.5-mini-3.8b","model_name":"Phi-3.5 mini 3.8B Instruct","params_b":3.79999999999999982236431605997495353221893310546875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":3.899999999999999911182158029987476766109466552734375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"phi4-14b","model_name":"Phi-4 14B Instruct","params_b":14.699999999999999289457264239899814128875732421875,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"deepseek-r1-distill-qwen-7b","model_name":"DeepSeek-R1-Distill-Qwen-7B","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3.29999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"deepseek-r1-distill-qwen-14b","model_name":"DeepSeek-R1-Distill-Qwen-14B","params_b":14,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"deepseek-r1-distill-qwen-32b","model_name":"DeepSeek-R1-Distill-Qwen-32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"deepseek-r1-distill-llama-70b","model_name":"DeepSeek-R1-Distill-Llama-70B","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"olmo2-7b","model_name":"OLMo 2 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"olmo2-13b","model_name":"OLMo 2 13B Instruct","params_b":13,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"command-r-35b","model_name":"Command R 35B (08-2024)","params_b":35,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rx-7600-8gb","gpu_name":"AMD Radeon RX 7600","gpu_vram_gb":8,"model_id":"command-r-plus-104b","model_name":"Command R+ 104B","params_b":104,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"smollm2-1.7b","model_name":"SmolLM2 1.7B Instruct","params_b":1.6999999999999999555910790149937383830547332763671875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":4.4000000000000003552713678800500929355621337890625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"llama3.2-1b","model_name":"Llama 3.2 1B Instruct","params_b":1,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"ample","recommended_quant":"fp16","headroom_at_recommended":5.79999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"llama3.2-3b","model_name":"Llama 3.2 3B Instruct","params_b":3,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"good","recommended_quant":"fp16","headroom_at_recommended":1.6999999999999999555910790149937383830547332763671875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"llama3.1-8b","model_name":"Llama 3.1 8B Instruct","params_b":8,"q4_fits":true,"q4_quality":"ample","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"llama3.1-70b","model_name":"Llama 3.1 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"llama3.3-70b","model_name":"Llama 3.3 70B Instruct","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"qwen2.5-7b","model_name":"Qwen2.5 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"qwen2.5-14b","model_name":"Qwen2.5 14B Instruct","params_b":14,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"qwen2.5-32b","model_name":"Qwen2.5 32B Instruct","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"qwen2.5-72b","model_name":"Qwen2.5 72B Instruct","params_b":72,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"qwen3-8b","model_name":"Qwen3 8B","params_b":8,"q4_fits":true,"q4_quality":"good","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":2.79999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"qwen3-14b","model_name":"Qwen3 14B","params_b":14,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"qwen3-32b","model_name":"Qwen3 32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"qwen3-30b-a3b","model_name":"Qwen3 30B-A3B (MoE)","params_b":30,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"qwen3-235b-a22b","model_name":"Qwen3 235B-A22B (MoE)","params_b":235,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"gemma3-4b","model_name":"Gemma 3 4B Instruct","params_b":4,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q8","headroom_at_recommended":3.70000000000000017763568394002504646778106689453125,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"gemma3-12b","model_name":"Gemma 3 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"tight","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":0.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"gemma3-27b","model_name":"Gemma 3 27B Instruct","params_b":27,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"gemma4-e4b","model_name":"Gemma 4 E4B (Edge)","params_b":4.5,"q4_fits":false,"q4_quality":"no","q8_fits":null,"q8_quality":"unknown","fp16_fits":null,"fp16_quality":"unknown","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"gemma4-31b","model_name":"Gemma 4 31B","params_b":31,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"mistral-7b-v0.3","model_name":"Mistral 7B v0.3 Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"mistral-nemo-12b","model_name":"Mistral Nemo 12B Instruct","params_b":12,"q4_fits":true,"q4_quality":"tight","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":0.90000000000000002220446049250313080847263336181640625,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"mixtral-8x7b","model_name":"Mixtral 8x7B Instruct v0.1","params_b":46.7000000000000028421709430404007434844970703125,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"phi3.5-mini-3.8b","model_name":"Phi-3.5 mini 3.8B Instruct","params_b":3.79999999999999982236431605997495353221893310546875,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"ample","fp16_fits":true,"fp16_quality":"tight","recommended_quant":"q8","headroom_at_recommended":3.899999999999999911182158029987476766109466552734375,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"phi4-14b","model_name":"Phi-4 14B Instruct","params_b":14.699999999999999289457264239899814128875732421875,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"deepseek-r1-distill-qwen-7b","model_name":"DeepSeek-R1-Distill-Qwen-7B","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3.29999999999999982236431605997495353221893310546875,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"deepseek-r1-distill-qwen-14b","model_name":"DeepSeek-R1-Distill-Qwen-14B","params_b":14,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"deepseek-r1-distill-qwen-32b","model_name":"DeepSeek-R1-Distill-Qwen-32B","params_b":32,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"deepseek-r1-distill-llama-70b","model_name":"DeepSeek-R1-Distill-Llama-70B","params_b":70,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"olmo2-7b","model_name":"OLMo 2 7B Instruct","params_b":7,"q4_fits":true,"q4_quality":"ample","q8_fits":true,"q8_quality":"tight","fp16_fits":false,"fp16_quality":"no","recommended_quant":"q4","headroom_at_recommended":3.5,"single_gpu":true,"multi_gpu_needed":false},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"olmo2-13b","model_name":"OLMo 2 13B Instruct","params_b":13,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"command-r-35b","model_name":"Command R 35B (08-2024)","params_b":35,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true},{"gpu_id":"rtx-3070-8gb","gpu_name":"NVIDIA GeForce RTX 3070","gpu_vram_gb":8,"model_id":"command-r-plus-104b","model_name":"Command R+ 104B","params_b":104,"q4_fits":false,"q4_quality":"no","q8_fits":false,"q8_quality":"no","fp16_fits":false,"fp16_quality":"no","recommended_quant":"none","headroom_at_recommended":null,"single_gpu":false,"multi_gpu_needed":true}]}