Z.ai's Open-Weights GLM-5.2 Beats GPT-5.5 on Long-Horizon Coding Benchmarks at One-Sixth the Cost

Summary

Chinese lab Z.ai (formerly Zhipu AI) released GLM-5.2, a 753B-parameter mixture-of-experts model (≈40B active params, 1.51TB) engineered for long-horizon autonomous coding and engineering tasks. Core weights shipped under an unrestricted MIT license on June 16 following a June 13 subscription launch, and the model carries a 1M-token context window (up from GLM-5.1's 200K). It is now the leading open-weights model on the Artificial Analysis Intelligence Index (51), scores 81.0 on Terminal-Bench 2