| { |
| "name": "ModelFit Local LLM Hardware Compatibility Dataset", |
| "description": "Which local AI models (Ollama) fit which hardware, by parameter size, quantization, and estimated memory load. tok/s and load figures are estimates, not measured benchmarks.", |
| "source": "https://modelfit.io/data/", |
| "license": "CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)", |
| "attribution": "ModelFit: https://modelfit.io/", |
| "methodology": { |
| "ramBudgetRatioBase": 0.7, |
| "ramBudgetRatioMax": 0.85, |
| "note": "A model fits a device when its estimatedLoadGb is within the memory budget: ~70% of unified memory up to 32GB, scaling linearly to ~85% at 128GB and above (high-RAM Macs can wire more memory to the GPU). Q4_K_M needs roughly 0.6 GB per billion parameters.", |
| "fieldSemantics": "runsLocally = a registry-verified Ollama build fits at least one consumer RAM tier tracked here (up to 256GB). openWeights = the weights are publicly downloadable. Some open-weight giants (e.g. NVIDIA Nemotron 3 Ultra, 550B, ~190GB at 2-bit) are openWeights: true but runsLocally: false because no consumer tier fits them." |
| }, |
| "updated": "2026-10-03", |
| "counts": { |
| "models": 156, |
| "local": 116, |
| "cloud": 40, |
| "families": 27 |
| }, |
| "models": [ |
| { |
| "model": "Qwen2.5 1.5B Instruct", |
| "family": "Qwen", |
| "params": 1.5, |
| "quantization": "Q4_K_M", |
| "minRamGb": 4, |
| "estimatedLoadGb": 1.5, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Translation", |
| "ollamaCommand": "ollama run qwen2.5:1.5b-instruct-q4_K_M" |
| }, |
| { |
| "model": "Qwen2.5 3B Instruct", |
| "family": "Qwen", |
| "params": 3, |
| "quantization": "Q4_K_M", |
| "minRamGb": 4, |
| "estimatedLoadGb": 2.5, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Coding", |
| "ollamaCommand": "ollama run qwen2.5:3b-instruct-q4_K_M" |
| }, |
| { |
| "model": "Qwen2.5 7B Instruct", |
| "family": "Qwen", |
| "params": 7, |
| "quantization": "Q4_K_M", |
| "minRamGb": 8, |
| "estimatedLoadGb": 5.5, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Coding", |
| "ollamaCommand": "ollama run qwen2.5:7b-instruct-q4_K_M" |
| }, |
| { |
| "model": "Qwen2.5 14B Instruct", |
| "family": "Qwen", |
| "params": 14, |
| "quantization": "Q4_K_M", |
| "minRamGb": 16, |
| "estimatedLoadGb": 11, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Coding, Chat", |
| "ollamaCommand": "ollama run qwen2.5:14b-instruct-q4_K_M" |
| }, |
| { |
| "model": "Llama 3.2 3B Instruct", |
| "family": "Llama", |
| "params": 3, |
| "quantization": "Q4_K_M", |
| "minRamGb": 4, |
| "estimatedLoadGb": 2.5, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat", |
| "ollamaCommand": "ollama run llama3.2:3b-instruct-q4_K_M" |
| }, |
| { |
| "model": "Llama 3.1 8B Instruct", |
| "family": "Llama", |
| "params": 8, |
| "quantization": "Q4_K_M", |
| "minRamGb": 12, |
| "estimatedLoadGb": 6.5, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Coding", |
| "ollamaCommand": "ollama run llama3.1:8b-instruct-q4_K_M" |
| }, |
| { |
| "model": "Llama 3.1 8B Instruct (Q8)", |
| "family": "Llama", |
| "params": 8, |
| "quantization": "Q8_0", |
| "minRamGb": 16, |
| "estimatedLoadGb": 8, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Coding", |
| "ollamaCommand": "ollama run llama3.1:8b-instruct-q8_0" |
| }, |
| { |
| "model": "Llama 3.1 8B Instruct (Q5)", |
| "family": "Llama", |
| "params": 8, |
| "quantization": "Q5_K_M", |
| "minRamGb": 12, |
| "estimatedLoadGb": 8, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Coding", |
| "ollamaCommand": "ollama run llama3.1:8b-instruct-q5_K_M" |
| }, |
| { |
| "model": "Llama 3.1 70B Instruct", |
| "family": "Llama", |
| "params": 70, |
| "quantization": "Q4_K_M", |
| "minRamGb": 64, |
| "estimatedLoadGb": 42, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Quality, Coding", |
| "ollamaCommand": "ollama run llama3.1:70b-instruct-q4_K_M" |
| }, |
| { |
| "model": "Mistral 7B Instruct", |
| "family": "Mistral", |
| "params": 7, |
| "quantization": "Q4_K_M", |
| "minRamGb": 8, |
| "estimatedLoadGb": 5.5, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Coding", |
| "ollamaCommand": "ollama run mistral:7b-instruct-q4_K_M" |
| }, |
| { |
| "model": "Mistral 7B Instruct (Q5)", |
| "family": "Mistral", |
| "params": 7, |
| "quantization": "Q5_K_M", |
| "minRamGb": 8, |
| "estimatedLoadGb": 4.8, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Coding", |
| "ollamaCommand": "ollama run mistral:7b-instruct-q5_K_M" |
| }, |
| { |
| "model": "Mistral 7B Instruct (Q8)", |
| "family": "Mistral", |
| "params": 7, |
| "quantization": "Q8_0", |
| "minRamGb": 16, |
| "estimatedLoadGb": 7.2, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Coding", |
| "ollamaCommand": "ollama run mistral:7b-instruct-q8_0" |
| }, |
| { |
| "model": "Mixtral 8x7B Instruct", |
| "family": "Mistral", |
| "params": 46.7, |
| "quantization": "Q4_K_M", |
| "minRamGb": 48, |
| "estimatedLoadGb": 30, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Coding, Quality", |
| "ollamaCommand": "ollama run mixtral:8x7b" |
| }, |
| { |
| "model": "Mistral Nemo 12B", |
| "family": "Mistral", |
| "params": 12, |
| "quantization": "Q4_K_M", |
| "minRamGb": 16, |
| "estimatedLoadGb": 9.5, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Translation", |
| "ollamaCommand": "ollama run mistral-nemo:12b" |
| }, |
| { |
| "model": "Mistral Nemo 12B (Q8)", |
| "family": "Mistral", |
| "params": 12, |
| "quantization": "Q8_0", |
| "minRamGb": 24, |
| "estimatedLoadGb": 12.1, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Translation", |
| "ollamaCommand": "ollama run mistral-nemo:12b-instruct-2407-q8_0" |
| }, |
| { |
| "model": "Gemma 2 2B Instruct", |
| "family": "Gemma", |
| "params": 2, |
| "quantization": "Q4_K_M", |
| "minRamGb": 4, |
| "estimatedLoadGb": 1.8, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat", |
| "ollamaCommand": "ollama run gemma2:2b-instruct-q4_K_M" |
| }, |
| { |
| "model": "Gemma 2 9B Instruct", |
| "family": "Gemma", |
| "params": 9, |
| "quantization": "Q4_K_M", |
| "minRamGb": 12, |
| "estimatedLoadGb": 7, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Coding", |
| "ollamaCommand": "ollama run gemma2:9b-instruct-q4_K_M" |
| }, |
| { |
| "model": "Gemma 2 27B Instruct", |
| "family": "Gemma", |
| "params": 27, |
| "quantization": "Q4_K_M", |
| "minRamGb": 32, |
| "estimatedLoadGb": 21, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Quality, Coding", |
| "ollamaCommand": "ollama run gemma2:27b-instruct-q4_K_M" |
| }, |
| { |
| "model": "Phi-3 Mini 3.8B", |
| "family": "Phi", |
| "params": 3.8, |
| "quantization": "Q4_K_M", |
| "minRamGb": 6, |
| "estimatedLoadGb": 3.2, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Coding, Chat", |
| "ollamaCommand": "ollama run phi3:mini" |
| }, |
| { |
| "model": "Phi-3 Medium 14B", |
| "family": "Phi", |
| "params": 14, |
| "quantization": "Q4_K_M", |
| "minRamGb": 16, |
| "estimatedLoadGb": 11, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Coding, Quality", |
| "ollamaCommand": "ollama run phi3:medium" |
| }, |
| { |
| "model": "Phi-4 14B", |
| "family": "Phi", |
| "params": 14, |
| "quantization": "Q4_K_M", |
| "minRamGb": 24, |
| "estimatedLoadGb": 11.5, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Coding, Quality", |
| "ollamaCommand": "ollama run phi4:14b-q4_K_M" |
| }, |
| { |
| "model": "Phi-4 14B (Q8)", |
| "family": "Phi", |
| "params": 14, |
| "quantization": "Q8_0", |
| "minRamGb": 24, |
| "estimatedLoadGb": 14.5, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Coding, Quality", |
| "ollamaCommand": "ollama run phi4:14b-q8_0" |
| }, |
| { |
| "model": "Qwen2.5 Coder 7B", |
| "family": "Qwen", |
| "params": 7, |
| "quantization": "Q4_K_M", |
| "minRamGb": 8, |
| "estimatedLoadGb": 5.5, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Coding", |
| "ollamaCommand": "ollama run qwen2.5-coder:7b" |
| }, |
| { |
| "model": "Qwen2.5 Coder 14B", |
| "family": "Qwen", |
| "params": 14, |
| "quantization": "Q4_K_M", |
| "minRamGb": 16, |
| "estimatedLoadGb": 11, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Coding", |
| "ollamaCommand": "ollama run qwen2.5-coder:14b" |
| }, |
| { |
| "model": "Llama 3.2 1B Instruct", |
| "family": "Llama", |
| "params": 1, |
| "quantization": "Q4_K_M", |
| "minRamGb": 2, |
| "estimatedLoadGb": 1, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat", |
| "ollamaCommand": "ollama run llama3.2:1b-instruct-q4_K_M" |
| }, |
| { |
| "model": "Mistral Small 22B", |
| "family": "Mistral", |
| "params": 22, |
| "quantization": "Q4_K_M", |
| "minRamGb": 32, |
| "estimatedLoadGb": 17, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Coding, Quality", |
| "ollamaCommand": "ollama run mistral-small:22b" |
| }, |
| { |
| "model": "Kimi K2 Instruct", |
| "family": "Kimi", |
| "params": 1000, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Quality, Coding", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "Claude 3.5 Sonnet", |
| "family": "Claude", |
| "params": null, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": false, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Coding, Chat", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "Claude 3.7 Sonnet", |
| "family": "Claude", |
| "params": null, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": false, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Coding, Quality", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "Claude 3 Opus", |
| "family": "Claude", |
| "params": null, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": false, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Quality, Analysis", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "Claude 4 Opus", |
| "family": "Claude", |
| "params": null, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": false, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Quality, Coding", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "Qwen2.5 0.5B Instruct", |
| "family": "Qwen", |
| "params": 0.5, |
| "quantization": "Q4_K_M", |
| "minRamGb": 2, |
| "estimatedLoadGb": 0.8, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Mobile", |
| "ollamaCommand": "ollama run qwen2.5:0.5b-instruct-q4_K_M" |
| }, |
| { |
| "model": "Gemma 3 1B Instruct", |
| "family": "Gemma", |
| "params": 1, |
| "quantization": "Q4_K_M", |
| "minRamGb": 2, |
| "estimatedLoadGb": 1, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Mobile", |
| "ollamaCommand": "ollama run gemma3:1b" |
| }, |
| { |
| "model": "Gemma 3 1B Instruct (Q8)", |
| "family": "Gemma", |
| "params": 1, |
| "quantization": "Q8_0", |
| "minRamGb": 4, |
| "estimatedLoadGb": 1, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Mobile", |
| "ollamaCommand": "ollama run gemma3:1b-it-q8_0" |
| }, |
| { |
| "model": "Phi-4 Mini 3.8B", |
| "family": "Phi", |
| "params": 3.8, |
| "quantization": "Q4_K_M", |
| "minRamGb": 6, |
| "estimatedLoadGb": 3.2, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Coding, Chat", |
| "ollamaCommand": "ollama run phi4-mini:3.8b" |
| }, |
| { |
| "model": "Phi-4 Mini 3.8B (Q8)", |
| "family": "Phi", |
| "params": 3.8, |
| "quantization": "Q8_0", |
| "minRamGb": 8, |
| "estimatedLoadGb": 3.8, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Coding, Chat", |
| "ollamaCommand": "ollama run phi4-mini:3.8b-q8_0" |
| }, |
| { |
| "model": "SmolLM2 360M", |
| "family": "SmolLM", |
| "params": 0.36, |
| "quantization": "Q4_K_M", |
| "minRamGb": 1, |
| "estimatedLoadGb": 0.5, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Embedded", |
| "ollamaCommand": "ollama run smollm2:360m" |
| }, |
| { |
| "model": "Llama 3.3 70B Instruct", |
| "family": "Llama", |
| "params": 70, |
| "quantization": "Q4_K_M", |
| "minRamGb": 64, |
| "estimatedLoadGb": 42, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Quality, Coding", |
| "ollamaCommand": "ollama run llama3.3:70b-instruct-q4_K_M" |
| }, |
| { |
| "model": "Llama 3.1 405B Instruct", |
| "family": "Llama", |
| "params": 405, |
| "quantization": "Q4_K_M", |
| "minRamGb": 320, |
| "estimatedLoadGb": 243, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Quality, Reasoning, Coding", |
| "ollamaCommand": "ollama run llama3.1:405b-instruct-q4_K_M" |
| }, |
| { |
| "model": "DeepSeek-R1 671B", |
| "family": "DeepSeek", |
| "params": 671, |
| "quantization": "Q4_K_M", |
| "minRamGb": 512, |
| "estimatedLoadGb": 380, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Reasoning, Coding", |
| "ollamaCommand": "ollama run deepseek-r1:671b-q4_K_M" |
| }, |
| { |
| "model": "Qwen3 8B", |
| "family": "Qwen", |
| "params": 8, |
| "quantization": "Q4_K_M", |
| "minRamGb": 12, |
| "estimatedLoadGb": 6.5, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Coding", |
| "ollamaCommand": "ollama run qwen3:8b-q4_K_M" |
| }, |
| { |
| "model": "Qwen3 8B (Q8)", |
| "family": "Qwen", |
| "params": 8, |
| "quantization": "Q8_0", |
| "minRamGb": 16, |
| "estimatedLoadGb": 8.1, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Coding", |
| "ollamaCommand": "ollama run qwen3:8b-q8_0" |
| }, |
| { |
| "model": "Qwen3 14B", |
| "family": "Qwen", |
| "params": 14, |
| "quantization": "Q4_K_M", |
| "minRamGb": 16, |
| "estimatedLoadGb": 11, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Coding, Quality", |
| "ollamaCommand": "ollama run qwen3:14b-q4_K_M" |
| }, |
| { |
| "model": "Qwen3 30B", |
| "family": "Qwen", |
| "params": 30, |
| "quantization": "Q4_K_M", |
| "minRamGb": 32, |
| "estimatedLoadGb": 22, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Quality, Coding", |
| "ollamaCommand": "ollama run qwen3:30b" |
| }, |
| { |
| "model": "Qwen3 30B (Q8)", |
| "family": "Qwen", |
| "params": 30, |
| "quantization": "Q8_0", |
| "minRamGb": 48, |
| "estimatedLoadGb": 30.3, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Quality, Coding", |
| "ollamaCommand": "ollama run qwen3:30b-a3b-q8_0" |
| }, |
| { |
| "model": "Gemma 3 4B Instruct", |
| "family": "Gemma", |
| "params": 4, |
| "quantization": "Q4_K_M", |
| "minRamGb": 6, |
| "estimatedLoadGb": 3.5, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Coding", |
| "ollamaCommand": "ollama run gemma3:4b" |
| }, |
| { |
| "model": "Gemma 3 4B Instruct (Q8)", |
| "family": "Gemma", |
| "params": 4, |
| "quantization": "Q8_0", |
| "minRamGb": 8, |
| "estimatedLoadGb": 3.9, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Coding", |
| "ollamaCommand": "ollama run gemma3:4b-it-q8_0" |
| }, |
| { |
| "model": "Gemma 3 12B Instruct", |
| "family": "Gemma", |
| "params": 12, |
| "quantization": "Q4_K_M", |
| "minRamGb": 16, |
| "estimatedLoadGb": 9.5, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Quality", |
| "ollamaCommand": "ollama run gemma3:12b" |
| }, |
| { |
| "model": "Gemma 3 27B Instruct", |
| "family": "Gemma", |
| "params": 27, |
| "quantization": "Q4_K_M", |
| "minRamGb": 32, |
| "estimatedLoadGb": 21, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Quality, Coding", |
| "ollamaCommand": "ollama run gemma3:27b" |
| }, |
| { |
| "model": "DeepSeek-R1 Distill Qwen 7B", |
| "family": "DeepSeek", |
| "params": 7, |
| "quantization": "Q4_K_M", |
| "minRamGb": 8, |
| "estimatedLoadGb": 5.5, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Reasoning, Coding", |
| "ollamaCommand": "ollama run deepseek-r1:7b" |
| }, |
| { |
| "model": "DeepSeek-R1 Distill Qwen 7B (Q8)", |
| "family": "DeepSeek", |
| "params": 7, |
| "quantization": "Q8_0", |
| "minRamGb": 16, |
| "estimatedLoadGb": 7.5, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Reasoning, Coding", |
| "ollamaCommand": "ollama run deepseek-r1:7b-qwen-distill-q8_0" |
| }, |
| { |
| "model": "DeepSeek-R1 Distill Qwen 14B", |
| "family": "DeepSeek", |
| "params": 14, |
| "quantization": "Q4_K_M", |
| "minRamGb": 16, |
| "estimatedLoadGb": 11, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Reasoning, Quality", |
| "ollamaCommand": "ollama run deepseek-r1:14b" |
| }, |
| { |
| "model": "DeepSeek-R1 Distill Qwen 14B (Q8)", |
| "family": "DeepSeek", |
| "params": 14, |
| "quantization": "Q8_0", |
| "minRamGb": 24, |
| "estimatedLoadGb": 14.6, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Reasoning, Quality", |
| "ollamaCommand": "ollama run deepseek-r1:14b-qwen-distill-q8_0" |
| }, |
| { |
| "model": "DeepSeek-R1 Distill Llama 70B", |
| "family": "DeepSeek", |
| "params": 70, |
| "quantization": "Q4_K_M", |
| "minRamGb": 64, |
| "estimatedLoadGb": 42, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Reasoning, Quality", |
| "ollamaCommand": "ollama run deepseek-r1:70b" |
| }, |
| { |
| "model": "GPT-4o", |
| "family": "OpenAI", |
| "params": null, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": false, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Quality, Coding", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "GPT-4o mini", |
| "family": "OpenAI", |
| "params": null, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": false, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Chat, Speed", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "Claude 4 Sonnet", |
| "family": "Claude", |
| "params": null, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": false, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Coding, Quality", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "Gemini 2.5 Pro", |
| "family": "Google", |
| "params": null, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": false, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Quality, Analysis", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "Gemini 2.5 Flash", |
| "family": "Google", |
| "params": null, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": false, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Speed, Chat", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "DeepSeek-V3", |
| "family": "DeepSeek", |
| "params": 671, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Quality, Coding", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "GLM-5", |
| "family": "Zhipu", |
| "params": 744, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Reasoning, Quality", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "GLM-4 Plus", |
| "family": "Zhipu", |
| "params": null, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": false, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Quality, Analysis", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "Qwen3 235B A22B", |
| "family": "Qwen", |
| "params": 235, |
| "quantization": "Q4_K_M", |
| "minRamGb": 192, |
| "estimatedLoadGb": 130, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Quality, Reasoning", |
| "ollamaCommand": "ollama run qwen3:235b-a22b-q4_K_M" |
| }, |
| { |
| "model": "Mistral Small 3.1", |
| "family": "Mistral", |
| "params": 24, |
| "quantization": "Q4_K_M", |
| "minRamGb": 24, |
| "estimatedLoadGb": 15, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Coding", |
| "ollamaCommand": "ollama run mistral-small3.1:24b" |
| }, |
| { |
| "model": "Mistral Small 3.1 (Q8)", |
| "family": "Mistral", |
| "params": 24, |
| "quantization": "Q8_0", |
| "minRamGb": 36, |
| "estimatedLoadGb": 23.3, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Coding", |
| "ollamaCommand": "ollama run mistral-small3.1:24b-instruct-2503-q8_0" |
| }, |
| { |
| "model": "DeepSeek-V3-0324", |
| "family": "DeepSeek", |
| "params": 671, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Quality, Coding", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "DeepSeek-R1", |
| "family": "DeepSeek", |
| "params": 671, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Reasoning, Quality", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "Qwen3.5 0.8B Instruct", |
| "family": "Qwen", |
| "params": 0.8, |
| "quantization": "Q4_K_M", |
| "minRamGb": 2, |
| "estimatedLoadGb": 0.8, |
| "kvKbPerToken": 12, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Mobile", |
| "ollamaCommand": "ollama run qwen3.5:0.8b" |
| }, |
| { |
| "model": "Qwen3.5 2B Instruct", |
| "family": "Qwen", |
| "params": 2, |
| "quantization": "Q4_K_M", |
| "minRamGb": 4, |
| "estimatedLoadGb": 1.8, |
| "kvKbPerToken": 12, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Edge tasks", |
| "ollamaCommand": "ollama run qwen3.5:2b" |
| }, |
| { |
| "model": "Qwen3.5 4B Instruct", |
| "family": "Qwen", |
| "params": 4, |
| "quantization": "Q4_K_M", |
| "minRamGb": 6, |
| "estimatedLoadGb": 3.5, |
| "kvKbPerToken": 32, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Coding, Agents, Multimodal", |
| "ollamaCommand": "ollama run qwen3.5:4b" |
| }, |
| { |
| "model": "Qwen3.5 4B Instruct (Q8)", |
| "family": "Qwen", |
| "params": 4, |
| "quantization": "Q8_0", |
| "minRamGb": 8, |
| "estimatedLoadGb": 4.3, |
| "kvKbPerToken": 32, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Coding, Agents, Multimodal", |
| "ollamaCommand": "ollama run qwen3.5:4b-q8_0" |
| }, |
| { |
| "model": "Qwen3.5 9B Instruct", |
| "family": "Qwen", |
| "params": 9, |
| "quantization": "Q4_K_M", |
| "minRamGb": 12, |
| "estimatedLoadGb": 7, |
| "kvKbPerToken": 32, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Quality, Coding, Reasoning", |
| "ollamaCommand": "ollama run qwen3.5:9b" |
| }, |
| { |
| "model": "Qwen3.5 35B-A3B Instruct", |
| "family": "Qwen", |
| "params": 35, |
| "quantization": "Q4_K_M", |
| "minRamGb": 32, |
| "estimatedLoadGb": 20, |
| "kvKbPerToken": 20, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Reasoning, Coding, Agent scenarios", |
| "ollamaCommand": "ollama run qwen3.5:35b-a3b" |
| }, |
| { |
| "model": "Qwen3.5 27B Instruct", |
| "family": "Qwen", |
| "params": 27, |
| "quantization": "Q4_K_M", |
| "minRamGb": 24, |
| "estimatedLoadGb": 16, |
| "kvKbPerToken": 64, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Coding, Complex reasoning", |
| "ollamaCommand": "ollama run qwen3.5:27b" |
| }, |
| { |
| "model": "Qwen3.5 27B Instruct (Q8)", |
| "family": "Qwen", |
| "params": 27, |
| "quantization": "Q8_0", |
| "minRamGb": 48, |
| "estimatedLoadGb": 27.1, |
| "kvKbPerToken": 64, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Coding, Complex reasoning", |
| "ollamaCommand": "ollama run qwen3.5:27b-q8_0" |
| }, |
| { |
| "model": "Qwen3.5 122B-A10B Instruct", |
| "family": "Qwen", |
| "params": 122, |
| "quantization": "Q4_K_M", |
| "minRamGb": 96, |
| "estimatedLoadGb": 72, |
| "kvKbPerToken": 24, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Frontier-level reasoning, Complex tasks", |
| "ollamaCommand": "ollama run qwen3.5:122b-a10b" |
| }, |
| { |
| "model": "LFM2 24B-A2B Instruct", |
| "family": "LFM2", |
| "params": 24, |
| "quantization": "Q4_K_M", |
| "minRamGb": 24, |
| "estimatedLoadGb": 14, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Local AI agents, privacy-first tool calling, MCP workflows", |
| "ollamaCommand": "ollama run lfm2:24b-a2b" |
| }, |
| { |
| "model": "LFM2.5 8B-A1B", |
| "family": "LFM2", |
| "params": 8.3, |
| "quantization": "Q4_K_M", |
| "minRamGb": 8, |
| "estimatedLoadGb": 5.5, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "On-device agents, tool calling, multilingual chat", |
| "ollamaCommand": "ollama run lfm2.5:8b-a1b-q4_K_M" |
| }, |
| { |
| "model": "Granite 4.1 3B Instruct", |
| "family": "Granite", |
| "params": 3, |
| "quantization": "Q4_K_M", |
| "minRamGb": 4, |
| "estimatedLoadGb": 2, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Lightweight chat, classification, edge tasks", |
| "ollamaCommand": "ollama run granite4.1:3b" |
| }, |
| { |
| "model": "Granite 4.1 8B Instruct", |
| "family": "Granite", |
| "params": 8, |
| "quantization": "Q4_K_M", |
| "minRamGb": 8, |
| "estimatedLoadGb": 5.5, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Enterprise assistant, tool calling, instruction following", |
| "ollamaCommand": "ollama run granite4.1:8b" |
| }, |
| { |
| "model": "Qwen3.6 27B", |
| "family": "Qwen", |
| "params": 27, |
| "quantization": "Q4_K_M", |
| "minRamGb": 32, |
| "estimatedLoadGb": 18, |
| "kvKbPerToken": 64, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Coding, Quality, Long context", |
| "ollamaCommand": "ollama run qwen3.6:27b" |
| }, |
| { |
| "model": "Qwen3.8 27B", |
| "family": "Qwen", |
| "params": 27, |
| "quantization": "Q4_K_M", |
| "minRamGb": 24, |
| "estimatedLoadGb": 16.5, |
| "kvKbPerToken": 64, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Coding, Agent, Vision, Long context", |
| "ollamaCommand": "ollama run qwen3.8:27b" |
| }, |
| { |
| "model": "Qwen3.8 27B (Q8)", |
| "family": "Qwen", |
| "params": 27, |
| "quantization": "Q8_0", |
| "minRamGb": 48, |
| "estimatedLoadGb": 27.1, |
| "kvKbPerToken": 64, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Coding, Agent, Vision, Long context", |
| "ollamaCommand": "ollama run qwen3.8:27b-q8_0" |
| }, |
| { |
| "model": "Qwen3.6 35B-A3B", |
| "family": "Qwen", |
| "params": 35, |
| "quantization": "Q4_K_M", |
| "minRamGb": 32, |
| "estimatedLoadGb": 22, |
| "kvKbPerToken": 20, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Reasoning, Coding, Agents", |
| "ollamaCommand": "ollama run qwen3.6:35b-a3b" |
| }, |
| { |
| "model": "Gemma 4 31B", |
| "family": "Gemma", |
| "params": 31, |
| "quantization": "Q4_K_M", |
| "minRamGb": 32, |
| "estimatedLoadGb": 20, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Quality, Coding, Multimodal", |
| "ollamaCommand": "ollama run gemma4:31b" |
| }, |
| { |
| "model": "Gemma 4 31B (Q8)", |
| "family": "Gemma", |
| "params": 31, |
| "quantization": "Q8_0", |
| "minRamGb": 48, |
| "estimatedLoadGb": 30.9, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Quality, Coding, Multimodal", |
| "ollamaCommand": "ollama run gemma4:31b-it-q8_0" |
| }, |
| { |
| "model": "Gemma 4 26B-A4B", |
| "family": "Gemma", |
| "params": 26, |
| "quantization": "Q4_K_M", |
| "minRamGb": 24, |
| "estimatedLoadGb": 16, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Coding, Multimodal", |
| "ollamaCommand": "ollama run gemma4:26b" |
| }, |
| { |
| "model": "Gemma 4 E4B", |
| "family": "Gemma", |
| "params": 4.5, |
| "quantization": "Q4_K_M", |
| "minRamGb": 6, |
| "estimatedLoadGb": 4, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "On-device, Mobile, Chat", |
| "ollamaCommand": "ollama run gemma4:e4b" |
| }, |
| { |
| "model": "Gemma 4 E4B (Q8)", |
| "family": "Gemma", |
| "params": 4.5, |
| "quantization": "Q8_0", |
| "minRamGb": 16, |
| "estimatedLoadGb": 7.5, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "On-device, Mobile, Chat", |
| "ollamaCommand": "ollama run gemma4:e4b-it-q8_0" |
| }, |
| { |
| "model": "Gemma 4 E2B", |
| "family": "Gemma", |
| "params": 2.3, |
| "quantization": "Q4_K_M", |
| "minRamGb": 4, |
| "estimatedLoadGb": 2.3, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "IoT, Mobile, Edge", |
| "ollamaCommand": "ollama run gemma4:e2b" |
| }, |
| { |
| "model": "Gemma 4 E2B (Q8)", |
| "family": "Gemma", |
| "params": 2.3, |
| "quantization": "Q8_0", |
| "minRamGb": 8, |
| "estimatedLoadGb": 4.6, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "IoT, Mobile, Edge", |
| "ollamaCommand": "ollama run gemma4:e2b-it-q8_0" |
| }, |
| { |
| "model": "Llama 4 Scout", |
| "family": "Llama", |
| "params": 109, |
| "quantization": "Q4_K_M", |
| "minRamGb": 96, |
| "estimatedLoadGb": 67, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Long context, Quality, Multimodal", |
| "ollamaCommand": "ollama run llama4:scout" |
| }, |
| { |
| "model": "Llama 4 Maverick", |
| "family": "Llama", |
| "params": 400, |
| "quantization": "Q4_K_M", |
| "minRamGb": 320, |
| "estimatedLoadGb": 245, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Frontier quality, Long context", |
| "ollamaCommand": "ollama run llama4:maverick" |
| }, |
| { |
| "model": "Mistral Medium 3.5", |
| "family": "Mistral", |
| "params": 128, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": true, |
| "ggufDiy": true, |
| "runtimes": [], |
| "bestFor": "Coding, Quality, Tool calling", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "DeepSeek V4 Flash 0731", |
| "family": "DeepSeek", |
| "params": 284, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": true, |
| "ggufDiy": true, |
| "runtimes": [], |
| "bestFor": "Reasoning, Coding, Long context", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "DeepSeek V4 Pro", |
| "family": "DeepSeek", |
| "params": 1600, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Frontier reasoning, Coding", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "DeepSeek V4.1 Flash", |
| "family": "DeepSeek", |
| "params": 552, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Agentic coding, Reasoning, Long context, Vision", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "Kimi K2.6", |
| "family": "Kimi", |
| "params": 1000, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Coding, Agentic tasks", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "GLM-5.1", |
| "family": "Zhipu", |
| "params": 744, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Coding, Reasoning, Long context", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "Claude Opus 4.7", |
| "family": "Claude", |
| "params": null, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": false, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Coding, Quality, Reasoning", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "Claude Opus 4.8", |
| "family": "Claude", |
| "params": null, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": false, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Coding, Agentic, Reasoning", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "GPT-5.5", |
| "family": "OpenAI", |
| "params": null, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": false, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Quality, Reasoning, Multimodal", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "Gemini 3.1 Pro", |
| "family": "Google", |
| "params": null, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": false, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Multimodal, Long context, Quality", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "Grok 4.3", |
| "family": "xAI", |
| "params": null, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": false, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Reasoning, Real-time data, Coding", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "Gemma 4 12B", |
| "family": "Gemma", |
| "params": 12, |
| "quantization": "Q4_K_M", |
| "minRamGb": 12, |
| "estimatedLoadGb": 8, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Coding, Multimodal", |
| "ollamaCommand": "ollama run gemma4:12b" |
| }, |
| { |
| "model": "Claude Fable 5", |
| "family": "Claude", |
| "params": null, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": false, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Coding, Agentic, Reasoning", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "GLM-5.2", |
| "family": "Zhipu", |
| "params": 753, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Coding, Reasoning, Agentic", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "Kimi K2.7-Code", |
| "family": "Kimi", |
| "params": 1000, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Coding, Agentic tasks", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "MiniMax M3", |
| "family": "MiniMax", |
| "params": 428, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Coding, Agentic, Long context", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "Qwen3.7-Plus", |
| "family": "Qwen", |
| "params": null, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": false, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Multimodal, Agentic, Vision", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "NVIDIA Nemotron 3 Ultra", |
| "family": "Nemotron", |
| "params": 550, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Reasoning, Agentic, Long context", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "Xiaomi MiMo-V2-Flash", |
| "family": "MiMo", |
| "params": 309, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": true, |
| "ggufDiy": true, |
| "runtimes": [], |
| "bestFor": "Reasoning, Coding, Agentic tasks", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "NVIDIA Nemotron Cascade 2 30B-A3B", |
| "family": "Nemotron", |
| "params": 30, |
| "quantization": "Q6_K", |
| "minRamGb": 36, |
| "estimatedLoadGb": 24, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Reasoning, Math, Agentic tasks", |
| "ollamaCommand": "ollama run nemotron-cascade-2:30b" |
| }, |
| { |
| "model": "Poolside Laguna XS.2", |
| "family": "Laguna", |
| "params": 33, |
| "quantization": "Q4_K_M", |
| "minRamGb": 36, |
| "estimatedLoadGb": 23, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Agentic coding, long-horizon software engineering", |
| "ollamaCommand": "ollama run laguna-xs.2:q4_K_M" |
| }, |
| { |
| "model": "Cohere North Mini Code", |
| "family": "North", |
| "params": 30, |
| "quantization": "Q4_K_M", |
| "minRamGb": 32, |
| "estimatedLoadGb": 19, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Agentic coding, repo-scale software engineering", |
| "ollamaCommand": "ollama run north-mini-code-1.0:q4_K_M" |
| }, |
| { |
| "model": "GPT-OSS 120B", |
| "family": "GPT-OSS", |
| "params": 117, |
| "quantization": "MXFP4", |
| "minRamGb": 96, |
| "estimatedLoadGb": 65.4, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Reasoning, Coding, Agents", |
| "ollamaCommand": "ollama run gpt-oss:120b" |
| }, |
| { |
| "model": "GPT-OSS 20B", |
| "family": "GPT-OSS", |
| "params": 21, |
| "quantization": "MXFP4", |
| "minRamGb": 24, |
| "estimatedLoadGb": 13.8, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Coding, Reasoning", |
| "ollamaCommand": "ollama run gpt-oss:20b" |
| }, |
| { |
| "model": "Qwen3-Next 80B-A3B", |
| "family": "Qwen", |
| "params": 80, |
| "quantization": "Q4_K_M", |
| "minRamGb": 72, |
| "estimatedLoadGb": 50.4, |
| "kvKbPerToken": 24, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Coding, Long Context", |
| "ollamaCommand": "ollama run qwen3-next:80b" |
| }, |
| { |
| "model": "Qwen3.6 35B-A3B (Q8)", |
| "family": "Qwen", |
| "params": 35, |
| "quantization": "Q8_0", |
| "minRamGb": 64, |
| "estimatedLoadGb": 38.7, |
| "kvKbPerToken": 20, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Reasoning, Coding, Agents", |
| "ollamaCommand": "ollama run qwen3.6:35b-a3b-q8_0" |
| }, |
| { |
| "model": "Qwen3.5 35B-A3B Instruct (Q8)", |
| "family": "Qwen", |
| "params": 35, |
| "quantization": "Q8_0", |
| "minRamGb": 64, |
| "estimatedLoadGb": 38.7, |
| "kvKbPerToken": 20, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Reasoning, Coding, Agent scenarios", |
| "ollamaCommand": "ollama run qwen3.5:35b-a3b-q8_0" |
| }, |
| { |
| "model": "Qwen3.5 9B Instruct (Q8)", |
| "family": "Qwen", |
| "params": 9, |
| "quantization": "Q8_0", |
| "minRamGb": 16, |
| "estimatedLoadGb": 10.7, |
| "kvKbPerToken": 32, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Quality, Coding, Reasoning", |
| "ollamaCommand": "ollama run qwen3.5:9b-q8_0" |
| }, |
| { |
| "model": "Qwen3.6 27B (Q8)", |
| "family": "Qwen", |
| "params": 27, |
| "quantization": "Q8_0", |
| "minRamGb": 48, |
| "estimatedLoadGb": 30, |
| "kvKbPerToken": 64, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Coding, Quality, Long context", |
| "ollamaCommand": "ollama run qwen3.6:27b-q8_0" |
| }, |
| { |
| "model": "Gemma 4 12B (Q8)", |
| "family": "Gemma", |
| "params": 12, |
| "quantization": "Q8_0", |
| "minRamGb": 24, |
| "estimatedLoadGb": 12.8, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Coding, Multimodal", |
| "ollamaCommand": "ollama run gemma4:12b-it-q8_0" |
| }, |
| { |
| "model": "Gemma 4 26B-A4B (Q8)", |
| "family": "Gemma", |
| "params": 26, |
| "quantization": "Q8_0", |
| "minRamGb": 48, |
| "estimatedLoadGb": 28.1, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Coding, Multimodal", |
| "ollamaCommand": "ollama run gemma4:26b-a4b-it-q8_0" |
| }, |
| { |
| "model": "Qwen3-Next 80B-A3B (Q8)", |
| "family": "Qwen", |
| "params": 80, |
| "quantization": "Q8_0", |
| "minRamGb": 128, |
| "estimatedLoadGb": 84.8, |
| "kvKbPerToken": 24, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Coding, Long Context", |
| "ollamaCommand": "ollama run qwen3-next:80b-a3b-instruct-q8_0" |
| }, |
| { |
| "model": "Qwen3 14B (Q8)", |
| "family": "Qwen", |
| "params": 14, |
| "quantization": "Q8_0", |
| "minRamGb": 24, |
| "estimatedLoadGb": 15.9, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Coding, Quality", |
| "ollamaCommand": "ollama run qwen3:14b-q8_0" |
| }, |
| { |
| "model": "Llama 3.3 70B Instruct (Q8)", |
| "family": "Llama", |
| "params": 70, |
| "quantization": "Q8_0", |
| "minRamGb": 96, |
| "estimatedLoadGb": 75, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Quality, Coding", |
| "ollamaCommand": "ollama run llama3.3:70b-instruct-q8_0" |
| }, |
| { |
| "model": "Llama 3.3 70B Instruct (Q6)", |
| "family": "Llama", |
| "params": 70, |
| "quantization": "Q6_K", |
| "minRamGb": 96, |
| "estimatedLoadGb": 57.9, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Quality, Coding", |
| "ollamaCommand": "ollama run llama3.3:70b-instruct-q6_K" |
| }, |
| { |
| "model": "Kimi K3", |
| "family": "Kimi", |
| "params": 2800, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Agents, Multimodal, Long context", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "Laguna XS 2.1", |
| "family": "Laguna", |
| "params": 33, |
| "quantization": "Q4_K_M", |
| "minRamGb": 32, |
| "estimatedLoadGb": 20.3, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Agentic coding, Long-horizon tasks", |
| "ollamaCommand": "ollama run laguna-xs-2.1:q4_K_M" |
| }, |
| { |
| "model": "Laguna S 2.1", |
| "family": "Laguna", |
| "params": 118, |
| "quantization": "Q4_K_M", |
| "minRamGb": 128, |
| "estimatedLoadGb": 96, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Agentic coding, Long-horizon tasks", |
| "ollamaCommand": "ollama run laguna-s-2.1:q4_K_M" |
| }, |
| { |
| "model": "Ornith 1.0 9B", |
| "family": "Ornith", |
| "params": 9, |
| "quantization": "Q4_K_M", |
| "minRamGb": 8, |
| "estimatedLoadGb": 5.6, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Agentic coding on small machines", |
| "ollamaCommand": "ollama run ornith:9b" |
| }, |
| { |
| "model": "Ornith 1.0 35B", |
| "family": "Ornith", |
| "params": 35, |
| "quantization": "Q4_K_M", |
| "minRamGb": 32, |
| "estimatedLoadGb": 21.2, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Agentic coding", |
| "ollamaCommand": "ollama run ornith:35b" |
| }, |
| { |
| "model": "Qwen3.8-Flash-Next", |
| "family": "Qwen", |
| "params": 125, |
| "quantization": "Q4_K_XL", |
| "minRamGb": 192, |
| "estimatedLoadGb": 111.3, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Agentic coding, Reasoning, Multimodal", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "Qwen3.8 2.4T A95B", |
| "family": "Qwen", |
| "params": 2400, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Frontier reasoning, Agentic coding", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "Granite 4.2 3B", |
| "family": "Granite", |
| "params": 3.7, |
| "quantization": "Q4_K_M", |
| "minRamGb": 4, |
| "estimatedLoadGb": 2.1, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Coding", |
| "ollamaCommand": "ollama run granite4.2:3b" |
| }, |
| { |
| "model": "Granite 4.2 8B", |
| "family": "Granite", |
| "params": 8.8, |
| "quantization": "Q4_K_M", |
| "minRamGb": 8, |
| "estimatedLoadGb": 5, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Coding", |
| "ollamaCommand": "ollama run granite4.2:8b" |
| }, |
| { |
| "model": "Granite 4.2 30B", |
| "family": "Granite", |
| "params": 29.3, |
| "quantization": "Q4_K_M", |
| "minRamGb": 24, |
| "estimatedLoadGb": 16.5, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Coding, Chat", |
| "ollamaCommand": "ollama run granite4.2:30b" |
| }, |
| { |
| "model": "Nemotron 3.5 Lightning 30B-A3B", |
| "family": "Nemotron", |
| "params": 30, |
| "quantization": "Q4_K_M", |
| "minRamGb": 36, |
| "estimatedLoadGb": 23.7, |
| "kvKbPerToken": 6, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Agentic, Coding, Long context", |
| "ollamaCommand": "ollama run nemotron-3.5-lightning:30b" |
| }, |
| { |
| "model": "Muse Glimmer 30B", |
| "family": "Muse", |
| "params": 29.8, |
| "quantization": "Q4_K_M", |
| "minRamGb": 32, |
| "estimatedLoadGb": 16.9, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat", |
| "ollamaCommand": "ollama run muse-glimmer:30b" |
| }, |
| { |
| "model": "MiniCPM-V 4.5 8B", |
| "family": "MiniCPM", |
| "params": 8.7, |
| "quantization": "Q4_K_M", |
| "minRamGb": 12, |
| "estimatedLoadGb": 5.7, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Vision, Multimodal", |
| "ollamaCommand": "ollama run minicpm-v4.5:8b" |
| }, |
| { |
| "model": "GLM-5.3", |
| "family": "Zhipu", |
| "params": 753, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Coding, Reasoning, Agentic", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "GLM-5.3 Flash", |
| "family": "Zhipu", |
| "params": 321, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": true, |
| "ggufDiy": true, |
| "runtimes": [], |
| "bestFor": "Chat, Coding", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "GLM-4.7-Flash 30B-A3B", |
| "family": "Zhipu", |
| "params": 30, |
| "quantization": "Q4_K_M", |
| "minRamGb": 32, |
| "estimatedLoadGb": 19, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Coding, Agentic tasks, Reasoning", |
| "ollamaCommand": "ollama run glm-4.7-flash" |
| }, |
| { |
| "model": "Nemotron 3 Super 120B-A12B", |
| "family": "Nemotron", |
| "params": 120, |
| "quantization": "Q4_K_M", |
| "minRamGb": 128, |
| "estimatedLoadGb": 86.8, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Agentic workflows, Long context, Reasoning", |
| "ollamaCommand": "ollama run nemotron-3-super:120b" |
| }, |
| { |
| "model": "Olmo 3.1 32B Instruct", |
| "family": "Olmo", |
| "params": 32, |
| "quantization": "Q4_K_M", |
| "minRamGb": 32, |
| "estimatedLoadGb": 19.5, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "General chat, Research, Fully open", |
| "ollamaCommand": "ollama run olmo-3.1:32b" |
| }, |
| { |
| "model": "MiMo-V2.6-Distill-Qwen-9B", |
| "family": "MiMo", |
| "params": 9, |
| "quantization": "Q4_K_M", |
| "minRamGb": 16, |
| "estimatedLoadGb": 5.8, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Agentic coding, Small machines", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "Ling-3.0-flash", |
| "family": "Ling", |
| "params": 124, |
| "quantization": "Q4_K_M", |
| "minRamGb": 128, |
| "estimatedLoadGb": 77.8, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Agentic coding, Long context, Reasoning", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "Xiaomi MiMo-V2.6-Flash", |
| "family": "MiMo", |
| "params": 309, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": true, |
| "ggufDiy": true, |
| "runtimes": [], |
| "bestFor": "Agentic coding, Multimodal, Long context", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "Xiaomi MiMo-V2.6-Pro", |
| "family": "MiMo", |
| "params": 1020, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [], |
| "bestFor": "Frontier agentic coding, Multimodal", |
| "ollamaCommand": null |
| }, |
| { |
| "model": "Devstral Small 2 24B", |
| "family": "Mistral", |
| "params": 24, |
| "quantization": "Q4_K_M", |
| "minRamGb": 24, |
| "estimatedLoadGb": 15.2, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Agentic coding, Software engineering", |
| "ollamaCommand": "ollama run devstral-small-2:24b" |
| }, |
| { |
| "model": "Devstral 2 123B", |
| "family": "Mistral", |
| "params": 123, |
| "quantization": "Q4_K_M", |
| "minRamGb": 96, |
| "estimatedLoadGb": 74.9, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Agentic coding, Software engineering", |
| "ollamaCommand": "ollama run devstral-2:123b" |
| }, |
| { |
| "model": "Ministral 3 3B", |
| "family": "Mistral", |
| "params": 3.4, |
| "quantization": "Q4_K_M", |
| "minRamGb": 6, |
| "estimatedLoadGb": 3, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Vision, Edge", |
| "ollamaCommand": "ollama run ministral-3:3b" |
| }, |
| { |
| "model": "Ministral 3 8B", |
| "family": "Mistral", |
| "params": 8.4, |
| "quantization": "Q4_K_M", |
| "minRamGb": 12, |
| "estimatedLoadGb": 6, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Coding, Vision", |
| "ollamaCommand": "ollama run ministral-3:8b" |
| }, |
| { |
| "model": "Ministral 3 14B", |
| "family": "Mistral", |
| "params": 13.5, |
| "quantization": "Q4_K_M", |
| "minRamGb": 16, |
| "estimatedLoadGb": 9.1, |
| "kvKbPerToken": null, |
| "runsLocally": true, |
| "openWeights": true, |
| "ggufDiy": false, |
| "runtimes": [ |
| "ollama", |
| "llama.cpp", |
| "lm-studio" |
| ], |
| "bestFor": "Chat, Reasoning, Vision", |
| "ollamaCommand": "ollama run ministral-3:14b" |
| }, |
| { |
| "model": "Inkling Small", |
| "family": "Inkling", |
| "params": 276, |
| "quantization": "API", |
| "minRamGb": 0, |
| "estimatedLoadGb": 0, |
| "kvKbPerToken": null, |
| "runsLocally": false, |
| "openWeights": true, |
| "ggufDiy": true, |
| "runtimes": [], |
| "bestFor": "Agentic coding, Multimodal, Reasoning", |
| "ollamaCommand": null |
| } |
| ] |
| } |
|
|
