GLM-5.2, an open-weight model from Z.ai, has demonstrated significant advancements, closing the gap with leading proprietary models. In just ten weeks, GLM-5.2 improved its Artificial Analysis Intelligence Index score by 11 points, reaching 51, which places it within five points of the top usable models like Claude Fable 5 and Claude Opus 4.8. The model shows substantial gains across various challenging evaluations, including physics reasoning, long-context reasoning, and agentic tasks, indicating broad capability improvements rather than a single benchmark trick. These advancements are attributed to architectural innovations like IndexShare, enhanced training techniques, and efficient serving methods, all while maintaining a similar model scale to its predecessor. AI
IMPACT Sets a new benchmark for open-weight models, significantly narrowing the performance gap with leading proprietary systems and accelerating research.
RANK_REASON Model release from a significant AI lab (Z.ai) with performance metrics competitive with top-tier proprietary models. [lever_c_demoted from frontier_release: ic=1 ai=1.0]
- Artificial Analysis
- Claude Fable 5
- Claude Opus 4.8
- GLM-5.2
- GPT-5.5
- Jie Tang
- Tsinghua University
- Wu Dao
- Z.ai
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
