Tencent Hunyuan multi-modal gets a new chief: Hu Han leaves to start a business, Tian Yonglong takes over VLM


According to Beating monitoring, Hu Han, head of Tencent Hunyuan’s multimodal understanding, has submitted his resignation and is preparing to launch a business. Former OpenAI researcher Tian Yonglong will take over the development of the vision language model (VLM), and will continue reporting to Yao Shunyu. Hu Han joined Tencent in early 2025 after coming from Microsoft Asia Research Institute. After that, he entered the Frontier research group in the Large Language Model department and participated in world model development. As Tian Yonglong takes over VLM, Hu Han’s original team may shift its focus toward world models. #​AI
View Original
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
  • Reward
  • 4
  • Repost
  • Share
Comment
Add a comment
Add a comment
TechLi
· 19h ago
A new CEO is always accompanied by uncertainty, but Tian Yonglong’s experience at OpenAI may bring fresh ideas to VLM. Hu Han’s entrepreneurship is also likely focused on a multimodal direction, and in the future it may compete with Tencent. Looking forward to further developments.
View OriginalReply0
SalvageCollector
· 20h ago
From Hu Hannan joining Tencent at the beginning of the year to leaving to start a business, it was only a little more than half a year—suggesting that internal R&D pressure was significant, or that the startup opportunity was more tempting. Tian Yonglong coming in from OpenAI to take over VLM indicates that Tencent is willing to poach talent, but the adjustments to the world model team are also worth watching—wondering whether it could affect the overall progress of Hunyuan’s multimodal work.
View OriginalReply0
RocketLauncher
· 20h ago
Tencent Hunyuan team’s roster change this round is pretty sudden. The VLM direction will be handed to Tian Yonglong, a former OpenAI employee. Presumably they want to speed up the race to catch up with Google and Meta’s vision-language models. Hopefully the team can stay stable.
View OriginalReply0
BTCRefugee
· 20h ago
It looks like the same kind of rhythm as someone like Hu Han—running off from Microsoft to Tencent and then starting a new venture again—suggesting that the big shots are all betting on the next big opportunity.
View OriginalReply0
  • Pinned