Tencent Reorganizes AI Research Units, Forms Unified Fundamental Model Department
Tencent recently announced a significant internal restructuring within its artificial intelligence research divisions. The company has decided to merge its Large Language Model (LLM) and Multimodal Model departments, forming a new, consolidated entity named the Fundamental Model Department. All associated projects and teams from both former units will be integrated into this new department.
Yao Shunyu Assumes Expanded Role, Leading Core Model Development
Under the new structure, Yao Shunyu, Tencent's Chief AI Scientist, will take on the additional role of heading the Fundamental Model Department, reporting directly to Lu Shan, President of the Technology and Engineering Group (TEG). This move significantly broadens Yao's responsibilities, placing him in direct charge of the company's most critical foundational model research and development efforts.
Tencent stated that the merger is primarily intended to enhance R&D efficiency and improve cross-team collaboration. By consolidating technical talent and resources previously spread across different departments, the company aims to pool its strengths more effectively to explore the technological frontiers of all-modal AI models.
Personnel Shifts Draw Attention, Future Strategy in Focus
The organizational changes have brought related personnel movements into the spotlight. The company's notice did not specify the future role of the former head of the multimodal model business. Furthermore, recent external reports suggest a leadership change in the multimodal understanding team, with a former OpenAI researcher reportedly stepping in. Tencent has not commented on these market rumors.
This restructuring clearly signals Tencent's effort to deeply integrate and focus its AI model development pipeline. Managing both language and multimodal capabilities under a single roof likely indicates a strategic push towards developing more general and powerful "all-modal" foundational models. The move appears to be a long-term play to build a more robust technological foundation in the highly competitive large model landscape.