colibri allows to run GLM-5.2 (744B-parameter MoE) on a consumer machine with
~25 GB of RAM - in pure C, with zero dependencies, by streaming experts from
disk.

Colibri is a lightweight, quality-preserving MoE runtime that treats VRAM, RAM,
and storage as one managed memory hierarchy. Insufficient fast memory may reduce
speed, but the default policy never silently changes model precision or router
semantics.
