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.
C avx2ccpu-inferenceggufinference-enginelinuxllmlocal-llmmixture-of-expertsmoenvmequantizationqwen3simdspeculative-decodingvulkan
| Name | Last commit message | Last commit date |
|---|---|---|
| .. | ||
| chat.md 30.7 KB | ||
| code-completion.md 7.4 KB | ||
| features.md 21.4 KB | ||
| hardware.md 11.4 KB | ||
| library.md 12.4 KB | ||
| models.md 15.5 KB | ||
| reports.md 5.3 KB | ||
| serving.md 39.5 KB | ||
| tools.md 3.9 KB | ||