Z.ai's GLM-5.2, a 744B-parameter open model with 40B active parameters and a 1M context window, can now run locally via Unsloth Dynamic GGUFs. Dynamic 2-bit quantization reduces disk space from 1.51TB to 239GB, fitting on a 256GB unified memory Mac. The model supports non-thinking and two thinking modes: High and Max.
Tap to vote and see what everyone thinks.
Summary by ByteBrief