LMSYS Teases Next GLM Model and Highlights GLM-5.2 Max Coding Performance
LMSYS Chatbot Arena has teased the upcoming release of the next GLM model version. Additionally, they highlighted that GLM-5.2 (Max) has achieved the #2 spot overall and #1 among open-source models in their Frontend Code Arena. This development showcases the rapid progress of open-source LLMs, with GLM-5.2 Max rivaling top closed-source models in coding capabilities. The upcoming GLM version could further narrow the gap or set new benchmarks for open-source AI performance. GLM-5.2 Max features a 1M-token context window and supports controllable reasoning effort. In benchmarks, it scored 81.0 on Terminal-Bench 2.1, placing it close to Claude 4.8 Opus and ahead of Gemini 3.1 Pro.
## BACKGROUND
LMSYS Chatbot Arena is a popular benchmarking platform that ranks large language models based on human preference and specific task arenas like coding. GLM (General Language Model) is a family of LLMs developed by Zhipu AI, known for strong performance in bilingual tasks and coding.