prabanta-dev/janas

37 tok/s from Qwen3-Next-80B (48 GB) on a 32 GB laptop, 24 without drafts vs 2.8 for llama.cpp; reading 62 vs 16 tok/s. Faster than llama.cpp on all 13 models compared in Q4_K_M. LLMs bigger than your RAM, pure C, CPU or GPU.

221 best points · 95 comments · 1 Hacker News post · 649 GitHub stars · View on GitHub

Hacker News discussions

• Janas: Run LLMs larger than RAM by streaming experts from SSD (221 points, 95 comments, by khalidkhair, 2026-10-12)