Local LLM database

30 open models worth running locally — tap any to see exactly which GPUs and Macs run it, and at what speed.

Data version 2026-06-22.

Edge & small (≤ 5B)

ModelParamsContextLicenseGood for
Llama 3.2 1B 1.2B 131K Llama 3.2 Community edge, general
Llama 3.2 3B 3.2B 131K Llama 3.2 Community edge, general, agents
Qwen3 4B 4B 131K Apache 2.0 edge, general, reasoning
Gemma 3 4B 4.3B 131K Gemma edge, vision, general

Consumer (5–20B)

ModelParamsContextLicenseGood for
Llama 3.1 8B 8.0B 131K Llama 3.1 Community general, agents, writing
Qwen3 8B 8.2B 131K Apache 2.0 general, reasoning, agents
Mistral 7B v0.3 7.3B 33K Apache 2.0 general, writing
Qwen2.5-Coder 7B 7.6B 131K Apache 2.0 coding, agents
Gemma 2 9B 9.2B 8K Gemma general, writing
Mistral Nemo 12B 12.2B 131K Apache 2.0 general, writing
Gemma 3 12B 12.2B 131K Gemma general, vision, writing
Phi-4 14B 14.7B 16K MIT reasoning, general
Qwen3 14B 14.8B 131K Apache 2.0 general, reasoning, coding
Qwen2.5 14B 14.8B 131K Apache 2.0 general, coding, writing
DeepSeek-R1 Distill Qwen 14B 14.8B 131K MIT reasoning, coding

Enthusiast (20–40B)

ModelParamsContextLicenseGood for
Mistral Small 3 24B 23.6B 33K Apache 2.0 general, coding, agents
Gemma 3 27B 27.4B 131K Gemma general, vision, writing
Gemma 2 27B 27.2B 8K Gemma general, writing
Qwen3 32B 32.8B 131K Apache 2.0 general, reasoning, coding
Qwen2.5 32B 32.8B 131K Apache 2.0 general, coding, writing
Qwen2.5-Coder 32B 32.8B 131K Apache 2.0 coding, agents
DeepSeek-R1 Distill Qwen 32B 32.8B 131K MIT reasoning, coding
Qwen3 30B-A3B (MoE)MoE 30.5B (3.3Ba) 131K Apache 2.0 general, reasoning, agents

Workstation & frontier (40B+)

ModelParamsContextLicenseGood for
Llama 3.3 70B 70.6B 131K Llama 3.3 Community general, coding, writing
Qwen2.5 72B 72.7B 131K Qwen general, coding, reasoning
DeepSeek-R1 Distill Llama 70B 70.6B 131K Llama 3.3 Community reasoning, coding
Mixtral 8x7B (MoE)MoE 46.7B (12.9Ba) 33K Apache 2.0 general
Llama 4 Scout (MoE)MoE 109B (17Ba) 1049K Llama 4 Community general, vision, agents
Qwen3 235B-A22B (MoE)MoE 235B (22Ba) 131K Apache 2.0 general, reasoning, coding
DeepSeek-R1 671B-A37B (MoE)MoE 671B (37Ba) 131K MIT reasoning, coding, general