GLM-5.2 Tops AI Ranking: Coding-First Model Shakes Up Open-Weights

Z.ai’s GLM-5.2 now leads the open-weights AI ranking, scoring 51 on Artificial Analysis’ Intelligence Index v4.1. This model offers public weights, a 1 million token context window, and API access. What's interesting is that GLM-5.2 is designed as a coding-first model. It achieved 62.1% on the SWE-bench Pro software-engineering benchmark. This compares to 58.6% for OpenAI’s GPT-5.5. The model also features a 128,000-token maximum output and an open-source MIT license for broad reuse. However, teams still need to consider hardware costs, API convenience, and data-routing requirements for deployment. The bottom line is that the availability of open weights makes GLM-5.2 a significant option for developers looking for flexible and cost-effective AI solutions.