TMTPOST — Chinese AI startup Zhipu AI has officially released and open-sourced its latest multimodal vision-language model, GLM-4.1V-Thinking, the company announced Tuesday. Designed for complex cognitive tasks, the general-purpose reasoning model supports multimodal inputs including images, video, and documents. Alongside the model launch, Zhipu unveiled a new ecosystem platform called “Agent Application Space” and kicked off its “Agents Pioneer Program,” committing hundreds of millions of yuan to support AI Agent startups.