{"source":"https://ossradar.net","license":"Free to use with a link back to ossradar.net","generatedAt":"2026-11-08T14:54:32.710Z","repo":{"fullName":"synchopate/llama-modchip","url":"https://github.com/synchopate/llama-modchip","ossradarUrl":"https://ossradar.net/repo/synchopate/llama-modchip","summary":"Run 284B DeepSeek-V4-Flash (<59GiB RAM, 44.4 tok/s pp, 9.1-14.7 tok/s) & 30B MoEs (25-111 tok/s) on a pure 64GB CPU at 256K ctx (0 GPUs).","description":"🐋 Run 284B DeepSeek-V4-Flash (<59GiB RAM, 44.4 tok/s pp, 9.1-14.7 tok/s) & 30B MoEs (25-111 tok/s) on a pure 64GB CPU at 256K ctx (0 GPUs). SwiGLU-H2, Stockfish-PVS L3 & 13ms Claude Code restore.","homepage":"https://github.com/synchopate/llama-modchip#readme","language":"Python","license":"Apache-2.0","category":"Local LLMs","categoryUrl":"https://ossradar.net/category/local-llms","topics":["ai-agents","apple-silicon","avx512","claude-code","cpu-inference","deepseek","deepseek-v4","gguf","hpc","llama-cpp","local-llm","mamba2","mixture-of-experts","nemotron","quantization","qwen3","speculative-decoding","stockfish","zen5"],"stars":189,"forks":18,"starsGained24h":101,"starsGained7d":138,"estimated":true,"rankToday":115,"rankThisWeek":449,"trackedRepos":450,"createdAt":"2026-10-29T10:17:33Z","pushedAt":"2026-11-08T14:28:05Z","starHistory":[{"date":"2026-11-02","stars":71},{"date":"2026-11-03","stars":72},{"date":"2026-11-04","stars":72},{"date":"2026-11-05","stars":73},{"date":"2026-11-06","stars":74},{"date":"2026-11-07","stars":88},{"date":"2026-11-08","stars":189}],"hackerNews":{"stories":1,"topStory":{"title":"Llama-modchip: 284B DeepSeek on 64GB RAM, no GPU","points":89,"comments":30,"date":"2026-11-08","url":"https://news.ycombinator.com/item?id=50411961"}},"alternativesUrl":"https://ossradar.net/alternatives/synchopate/llama-modchip","alternatives":["mw00/project-maya","nace-ai/kapitsa","syv-ai/HyperQwen","simplexai-labs/PolyStrata","RustingStudio/RustingStrata"],"badgeUrl":"https://ossradar.net/badge/synchopate/llama-modchip"}}