Fast (20+ tok/s)

Qwen3.5 0.8B 96DeepSeek-R1 1.5B 87Qwen 1.8B 87lfm 2.5 8B (A1B) 78Llama 3.2 1B 74Qwen 3 1.7B 69Granite 3.1 moe 1B 69minicpm v4.6 1B 60Gemma 2 2B 60codeGemma 2B 60Gemma 2B 57smollm 2 1.7B 53starCoder 2 3B 53falcon 3 1B 53Qwen2.5 3B 51Qwen2.5 Coder 3B 51Qwen3.5 2B 51Qwen 3-VL 2B 51Llama 3.2 3B 48orca mini 3B 48Granite 3.1 moe 3B 48falcon 3 3B 48Granite 4.1 3B 46Granite 4.2 3B 44Phi 3 3.8B 44Qwen 4B 42Qwen 3 4B 37Qwen2.5-VL 3B 30Gemma 3 4B 29Qwen 3-VL 4B 29translateGemma 4B 29Qwen3.5 4B 28Llama 2 7B 25Code Llama 7B 25DeepSeek Coder 6.7B 25Llama 2 uncensored 7B 25orca mini 7B 23Mistral 7B 22Qwen2 7B 22starCoder 2 7B 22OLMo 2 7B 21falcon 3 7B 21DeepSeek-R1 7B 20Qwen2.5 7B 20Llama 3 8B 20Qwen2.5 Coder 7B 20LLaVA 7B 20

Comfortable (10-20 tok/s)

Llama 3.1 8B 20Qwen 7B 20dolPhin 3 8B 20Gemma 7B 19codeGemma 7B 19DeepSeek-R1 8B 19Qwen 3 8B 19Granite 4.2 8B 18Granite 4.1 8B 18Gemma 2 9B 18minicpm v 8B 18LLaVA Llama3 8B 18ornith 9B 17Gemma 3n E2B 17Qwen2.5-VL 7B 16minicpm v4.5 8B 16Qwen 3-VL 8B 16falcon 3 10B 15ornith 1.5 9B 15Qwen3.5 9B 15Mistral nemo 12B 14Gemma 4 E2B 13Llama 2 13B 13Code Llama 13B 13Gemma 3n E4B 13Gemma 4 12B 13Llama 3.2 Vision 11B 12Phi 3 14B 12orca mini 13B 12LLaVA 13B 12Gemma 3 12B 12translateGemma 12B 12Qwen 14B 12OLMo 2 13B 11DeepSeek Coder v2 16B 11DeepSeek-R1 14B 11Qwen2.5 14B 11Qwen2.5 Coder 14B 11Phi 4 14B 11

Slow but usable (5-10 tok/s)

Qwen 3 14B 6.2Gemma 4 E4B 6.0starCoder 2 15B 5.8

Won't fit

Mistral small 22Bgpt-oss 20BMistral small 24BMistral small3.2 24BGemma 2 27BGranite 4.1 30BGemma 3 27BQwen3.5 27BQwen3.6 27BtranslateGemma 27BGranite 4.2 30BQwen3.8 27Bmuse glimmer 30BQwen 32BQwen 3 30B (A3B)Gemma 4 26B (A4B)Qwen3 Coder 30B (A3B)Code Llama 34BDeepSeek Coder 33BDeepSeek-R1 32BQwen2.5 32BQwen 3 32BGemma 4 31BQwen2.5 Coder 32BLLaVA 34BQwen 3-VL 30B (A3B)qwq 32Bornith 35BQwen 3-VL 32BQwen2.5-VL 32Bornith 1.5 35BQwen3.5 35B (A3B)Qwen3.6 35B (A3B)Nemotron 3.5 lightning 30B (A3B)mixtral 8X7BNemotron 3 33BLlama 2 70BCode Llama 70Borca mini 70BLlama 2 uncensored 70BLlama 3 70BQwen 72BQwen2 72BLlama 3.1 70BDeepSeek-R1 70BLlama 3.3 70BQwen2.5 72BQwen2.5-VL 72BLlama 3.2 Vision 90BQwen 110Bgpt-oss 120BMistral medium 3.5 128Bmixtral 8X22BQwen3.5 122B (A10B)Nemotron 3 super 120B (A12B)Qwen3.8 flash next 125B (A6B)