GLM-5.2 extends to 1M context and surpasses GPT-5.5 on engineering benchmarks at 1/6 the cost

Z.ai (Zhipu AI)/GLM-5.2

Z.ai released GLM-5.2, a 753-billion-parameter open-weight model optimized for long-horizon coding and reasoning. The model supports 1-million-token context via IndexShare sparse attention (2.9× compute reduction), scores 81.0 on Terminal-Bench 2.1 and 62.1 on SWE-bench Pro, and reportedly edges past GPT-5.5 on multi-step engineering tasks while costing roughly one-sixth as much. Released under MIT license on Hugging Face.