Zhipu: Visual reasoning model GLM-4.5V officially launched and open sourced.

Zhipu announced the release of the open-source visual reasoning model GLM-4.5V, which is also open-sourced on the Moda Community and Hugging Face. According to the introduction, GLM-4.5V is based on Zhipu's next-generation flagship text-based model GLM-4.5-Air, following the GLM-4.1V-Thinking technology route. It achieves comparable performance to SOTA in 41 public visual multimodal rankings, covering common tasks such as image, video, document understanding, and GUI agents.