Global large model "New Three Musketeers" pattern, Zhipu's (02513) latest programming capability approaches the strongest overseas.

According to the CBN APP, Zhipu (02513) released and open sourced the new flagship large model GLM-5.2 on June 17. In the Artificial Analysis comprehensive list, GLM-5.2 scored 51 points, ranking among the SOTA open source models. According to the list, top model manufacturers such as Anthropic, OpenAI, and Zhipu have formed a "new three major families" pattern. In the Code Arena code evaluation system with over a million blind users, GLM-5.2 ranked first in globally available models. GLM-5.2 added 1 million long context and long-range task capabilities. In authoritative benchmarks such as FrontierSWE and Terminal-Bench, GLM-5.2 narrowed the gap with the international top model Claude Opus 4.8 to 1%4%.