GLM-5.2 has emerged as a leading open-weight AI model, securing the top position in multiple benchmark categories. It achieved a remarkable ~1360 Elo in the Design Arena, surpassing Fable 5 by approximately 10 Elo points. Additionally, GLM-5.2 became the first open-weight model to exceed 80% on Terminal-Bench 2.1, with a best run of 82.7%. The model also leads the Artificial Analysis Intelligence Index v4.1 with a score of 51 and is competitive with, or ahead of, GPT-5.5 in GDPval-AA v2. In LiveBench Agentic Coding, it ranks #1–2 among open models. GLM-5.2 holds top-3 positions in several other benchmarks, including FrontierSWE, SWE-bench Pro, and MCP-Atlas, showcasing its comprehensive capabilities across various AI tasks.