Z.ai GLM-5.2: 1M-Token Context for Coding Tasks

Z.ai has released GLM-5.2, a new flagship model with a 1-million-token context. This model is designed for long-horizon coding tasks. What's interesting is GLM-5.2 has 753 billion parameters and introduces an optimization called IndexShare. This reduces per-token FLOPs by 2.9 times at the 1M context length. Z.ai published the core weights under an MIT license, allowing unrestricted commercial use. The model is available to Coding Plan subscribers and saw wider releases and benchmarks around June 16. On the Terminal-Bench 2.1, GLM-5.2 scored 81.0, compared to 85.0 for Claude Opus 4.8. This shows it challenges proprietary models in long-horizon coding. The bottom line is that open-weight models like GLM-5.2 offer practical options for businesses needing local hosting or customization.