Chinese AI lab Z.ai released GLM-5.2, a 753B-parameter text-only open-weights LLM under MIT license. It achieved the highest score among open-weights models on the Artificial Analysis Intelligence Index and supports a 1 million token context window.
Chinese AI lab Z.ai released GLM-5.2, a 753B-parameter (40B active, MoE) text-only open-weights LLM under MIT license. It supports a 1 million token context window, up from 200k in GLM-5.1.
GLM-5.2 was first released to coding plan subscribers on June 13, then fully open-weights on June 16. It scored 51 on the Artificial Analysis Intelligence Index v4.1, leading open-weights models ahead of MiniMax-M3 (44), DeepSeek V4 Pro (44), and Kimi K2.6 (43). However, it uses more output tokens per task (43k) compared to competitors.
Despite being text-only, GLM-5.2 ranks 2nd on the Code Arena WebDev leaderboard, demonstrating strong front-end coding capabilities without image input. The continued improvement of open-weights models narrows the gap with closed models and promotes AI democratization.
Some comments question the sharp performance drop of GLM-5.2 on a trivial task compared to its predecessor, suggesting possible benchmark contamination. Others note that models like Minimax M3 outperform GLM-5.2 yet receive less press, criticizing the hype-driven coverage that overlooks strong contenders.