arizqi/cpubrrr

Frontier-class LLM inference on a laptop CPU — gpt-oss:20b at ~110 tok/s on Apple M4 Max, 7.5x llama.cpp, no GPU. From-scratch NEON/SME kernels in Rust.

1 best points · 0 comments · 1 Hacker News post · 57 GitHub stars · View on GitHub

Hacker News discussions

• Frontier-class LLM inference on a laptop CPU (1 points, 0 comments, by yogthos, 2026-07-23)