Back to Featured
Simon Willison’s Blog2026-06-18 07:58June 18

GLM-5.2: Possibly the Most Powerful Open-Weight Large Language Model for Text Only

GLM-5.2 is probably the most powerful text-only open weights LLM

AI Briefing

Zhipu (Z.ai) released GLM-5.2 to coding-plan subscribers on June 13 and open-sourced its complete weights under the MIT License on June 16. The model has 753B parameters and is a 1.51TB MoE architecture with 40 active parameters. It accepts text-only input and has a context window expanded to 1 million tokens. It leads the Artificial Analysis Intelligence Index v4.1 with a score of 51, surpassing MiniMax-M3 (44), DeepSeek V4 Pro (max, 44), and Kimi K2.6 (43). However, its output-token consumption per task reaches 43k, higher than comparable models. It ranks second on the Code Arena WebDev leaderboard, behind only Claude Fable 5. It is available through OpenRouter, with most providers charging $1.40 per million input tokens and $4.40 per million output tokens. In testing, it performed excellently at generating SVG animations, but the quality of its opossum SVGs was inferior to GLM-5.1.

Also covered by 7 other sources

Source: Simon Willison's blog · simonwillison.net

Finished reading the news? Let’s talk about your business.

Leave your contact details, and I’ll get in touch with you personally.

WECHAT

Add me on WeChat

Press and hold or select the WeChat ID below to copy it.