en.Wedoany.com Reported - On May 15, Baidu officially established the Baidu Model Committee (BMC). Composed of young researchers with deep understanding of large model technology, the committee will be uniformly responsible for the entire process of Baidu's large models, from technological path exploration and planning, to model training and iteration, and finally to business scenario implementation. The original Basic Model R&D Unit (BMU) and Applied Model R&D Unit (AMU) will both report to the BMC.
This move signifies that Baidu has elevated the decision-making power and resource allocation authority for large models from decentralized business lines to the group's highest technical decision-making level. The newly established BMC is not merely an R&D department, but a committee entity with overall planning and technical management functions. Its goal is to break down the barriers between the basic model R&D team and the business application team, achieving a closed loop of "model integration from technology to application." The establishment of the BMC means that all path exploration, technical planning, model training, and business implementation related to large models within Baidu are now incorporated into a unified technical review and decision-making process.
The formation of the BMC is the latest move in a series of adjustments to Baidu's core AI organizational structure since November 2025. On November 25, 2025, Baidu announced the creation of the Basic Model R&D Unit (BMU) and the Applied Model R&D Unit (AMU), led by Wu Tian and Jia Lei respectively, with both departments reporting directly to CEO Robin Li. CTO Wang Haifeng continued to serve as Chairman of the Technology Strategy Committee (TSC) and President of Baidu Research, coordinating the overall technical layout. The adjustment at that time addressed the issue of advancing large model R&D and application in parallel, while the establishment of the BMC now further integrates the bifurcated dual engines under a unified decision-making framework.
At the talent level, the establishment of the BMC echoes Baidu's recently vigorously promoted strategy of recruiting top AI talent. Baidu previously announced that its AIDU (AI Leadership Talents) program, targeting top global campus recruits for technical positions, has been upgraded to a group-level talent development project. The recruitment quota for the 2026 cohort has expanded by over 60%, covering 23 core businesses and 11 research directions including large model algorithms, heterogeneous computing, and multimodality, with a compensation package that "has no upper limit." In May 2026, Baidu simultaneously launched the "Wenxin New Star" program, personally guided by CTO Wang Haifeng, focusing on young talents with disruptive innovative thinking. In connection with the recent trend of several well-known large model R&D talents in the industry joining Baidu, it is expected that the newly established BMC will have a significant number of recently joined young technical backbones taking committee seats.
The position of the BMC within Baidu's technical system has formed a clear division of decision-making responsibilities with the TSC led by CTO Wang Haifeng. The TSC continues to lead overall technical strategy and basic research, while the BMC focuses on the engineering implementation and full lifecycle management of large models. This architecture divides the R&D system into two logical layers: the Basic Model R&D Unit (BMU) focuses on frontier exploration and general capability building of large models, the Applied Model R&D Unit (AMU) specializes in fine-tuning dedicated models for industries such as finance, healthcare, and education, while the BMC, as the highest decision-making committee at the model level, exercises unified jurisdiction over technical route selection, resource allocation, and version release cadence.
Baidu is undergoing a shift in focus within the AI track from a "model arms race" to "application value delivery." At the Create 2026 Baidu AI Developer Conference, which opened on May 13, Baidu founder Robin Li explicitly proposed that the yardstick of the AI era is shifting from Tokens to DAA (Daily Active Agents), emphasizing that AI competition has moved from intelligence to execution capability. Baidu AI Cloud concurrently released a full-stack strategy upgrade of "Chip, Cloud, Model, Agent," launching over 30 new capabilities around the three major directions of AI Infra, Agent Infra, and AI Agent. The establishment of the BMC serves precisely as the supporting organizational-level technical decision-making hub for this strategic transformation—ensuring highly coordinated full-chain technical decisions from underlying chip adaptation (Kunlun Core), cloud infrastructure, and the Wenxin large model, up to upper-layer agent applications.
Baidu's establishment of the Model Committee this time keeps pace with the global trend among tech giants of organizational restructuring for the AI era. By centralizing the decision-making power for model R&D into the BMC, Baidu aims to maintain organizational agility through a younger, technically-oriented decision-making team, to meet the demands of rapid technological iteration from model training to application deployment.
This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com
