Hunyuan-Vision-1.5-Thinking arrives on Tencent Cloud, weights promised

Tencent's October 2025 announcement made Hunyuan-Vision-1.5-Thinking available through its cloud service for tasks that combine pictures and written questions. The project documentation shows an application submitting an image and a text prompt together, allowing the model to answer questions about visual content.

At that announcement, downloadable weights and a technical report were still promised for later release. The documentation also marked the separate mode for manipulating images during reasoning as forthcoming. The immediately described access was therefore a hosted vision-language service.