Blog

Janas-LLM outperforms llama.cpp on 13 models running larger-than-RAM LLMs on a 32 GB laptop

Janas-LLM runs Qwen3-Next-80B at 37 tokens per second on a 32 GB laptop, outperforming llama.cpp across 13 models while streaming experts from NVMe SSD on a Debian 13 system with Intel Core Ultra 9 185H.