Architecture Digest

Sep 15, 2026 · Artificial Intelligence

Running 744B GLM-5.2 on 25GB RAM: colibri's Disk-Streaming MoE Inference Engine

colibri, a zero-dependency C inference engine, runs Zhipu's 744B-parameter GLM-5.2 MoE model on a 25GB RAM laptop by keeping dense weights in memory and streaming routed experts from NVMe, achieving ~1 token/s after cache warm-up with router-lookahead prefetching and learned caching.