Exclusive｜Hu Han, the head of Hunyuan Multimodal Understanding, resigned to start a business. The original team may focus on world models

Text | Zhou Xinyu Editor | Zhang Yuxin "Smart Emergence" has learned exclusively that recently, Hu Han, the head of Tencent's Hunyuan multimodal understanding, proposed to resign. Previously, he served as the principal researcher of the Visual Computing Group at Microsoft Research Asia. After joining Tencent in early 2025, he was responsible for the research of large visual models. In the subsequent adjustments, he joined the "Frontier" cutting-edge technology research group under the large language model department, responsible for related research on multi-modal understanding, and reported to Yao Shunyu. It is understood that Hu Han has also been responsible for the research and development of world models. At the same time, Yao Shunyu, head of Tencent's large language model department, is intensively sorting out his teams recently. Hu Han's previous research group may focus on cutting-edge research on world models. As of press time, Tencent has not responded to the above information. Large language models are still the most important thing for Hunyuan. Since mid-2026, personnel changes surrounding the Hunyuan system have been ongoing. In early July, Tian Yonglong, former OpenAI Research Institute and MIT Ph.D., joined the large language model department - we learned that he will replace Hu Han in the research and development of the visual language model (VLM) and report to Yao Shunyu. Since Yao Shunyu joined, he has made up for the shortcomings and quickly improved the capabilities of the Hunyuan base model to the first echelon, which has become the main line of Tencent AI. Reorganizing resource allocation and focusing talents and computing power on the basic model research and development responsible for the large language model department is one of Yao Shunyu's core initiatives after taking office. For example, after Tencent canceled its AI Lab, all core R&D personnel were merged into the large language model department. The recruitment of large language model talents has also increased with the addition of Yao Shunyu, such as several people from the Byte Seed Infra team and post-training team. Later, at the Tencent Cloud AI Industry Application Conference on June 5, Tang Daosheng, senior executive vice president of Tencent Group and CEO of the Cloud and Smart Industry Group, asked Yao Shunyu a sharp question for the outside world to some extent: Many people say that Tencent is slow in AI. Do you think we are really slow? But in fact, "Shunyu is more anxious than Tencent's top management." A Hunyuan researcher told "Intelligent