Intelligent Chart: GLM-5.2 is online and open source today.

Zhipu announced today that GLM-5.2 has been launched and open-sourced. On the front-end development evaluation system CodeArena, where millions of users worldwide participated in blind testing, GLM-5.2 achieved the best performance among global available models. GLM-5.2 is specifically designed for long-range task capabilities, with new features including Solid1M context for stable support of long-range tasks, enhanced sensory perception, more practical coding abilities, ultimate Infra optimization running on domestic computing platforms from Day 0, MIT open source license without geographical restrictions promoting technical equality across borders. The 1M context constitutes GLM-5.2's long-range delivery capabilities, with multiple long-range task benchmarks indicating performance between Claude Opus 4.7 and 4.8, making it the highest-ranked open-source model. GLM-5.2's online inference relies on multiple domestic computing platforms and has achieved inference compatibility with domestic computing platforms such as Huawei Ascend, Pingtouge, Moore Threads, Cambricon, Kunlun, Muxi, Haiguang, and Biren on Day 0, ensuring stable operation with high throughput, low latency, and large concurrency on domestic chip clusters. After the launch of the Ascend 950 super node in the second half of the year, it will also become the powerful computing base for GLM-5.2.