GLM-5.2 is probably the most powerful text-only open weights LLM
Chinese AI lab Z.ai released GLM-5.2 to coding plan subscribers on June 13, 2026, and then on June 16 released the full open weights under an MIT license. This is a 753B parameter, 1.51TB model with 40 active parameters (Mixture of Experts). It is text-only, with a separate vision family (GLM-5V-Turbo) that is not open weights. The context window is 1 million tokens, up from GLM-5.1's 200,000. According to Artificial Analysis, GLM-5.2 is the leading open weights model on the Intelligence Index v4.1 with a score of 51, ahead of MiniMax-M3 (44), DeepSeek V4 Pro (max, 44), and Kimi K2.6 (43). However, it is token-hungry, using 43k output tokens per Intelligence Index task, compared to GLM-5.1 (26k), MiniMax-M3 (24k), Kimi K2.6 (35k), and DeepSeek V4 Pro (max, 37k). The model ranks 2nd on the Code Arena WebDev leaderboard, behind only Claude Fable 5, which measures front-end web development tasks including agentic coding workflows. Pricing via OpenRouter is around $1.40/million input tokens and $4.40/million output tokens from nine providers. The author tested the model's SVG generation: a pelican riding a bicycle produced a well-animated SVG, but a North Virginia opossum on an e-scooter was a step down from GLM-5.1's version.
GLM-5.2 offers a powerful open weights alternative for text-only tasks, leading benchmarks and rivaling proprietary models.