At a critical stage where the competition in large models is shifting from isolated breakthroughs to systematic operations, Baidu has recently announced the establishment of the **Baidu Model Committee (BMC)**, as the highest decision-making and coordination body for the company's large model strategy. The committee is composed of a group of **young researchers** with profound understanding of large model technologies. The Basic Model Research Department (BMU) and the Applied Model Research Department (AMU) will directly report to it, achieving integrated management throughout the entire chain from underlying technology development to upper-level product deployment.

This organizational restructuring marks Baidu's approach of "concentrating strengths to win decisive battles" in response to increasingly fierce AI competition. In the past, basic model teams and application teams often faced challenges in collaboration due to differing objectives — the former pursued SOTA metrics, while the latter focused on user experience and commercialization. The establishment of BMC aims to break down this barrier, ensuring that technological advancements and product demands are in sync.

According to insiders, although BMC members are mainly young scientists, they have all been deeply involved in multiple iterations of the ERNIE large model, possessing comprehensive experience in algorithms, engineering, and productization. They will be responsible for setting the technical roadmap for large models, prioritizing resource allocation, establishing cross-departmental collaboration mechanisms, and reviewing key projects, truly promoting the integration of "technology-product-commerce."

This move also represents a new strategic upgrade following the monthly active user count of ERNIE Bot exceeding 200 million. As AI enters the phase of "deep scenario cultivation," simply increasing parameters is no longer sufficient to build a moat. An **efficient closed-loop organizational capability** has become the key to success. Through BMC, Baidu hopes to accelerate the realization of an agile development model of "one model, multi-terminal reuse, and rapid iteration." For example, real feedback from business areas such as search, maps, and health can be fed back into model optimization in real time.

Amidst the global tech giants reorganizing their AI structures (such as Google integrating DeepMind or Microsoft strengthening the Copilot engineering system), Baidu's move is both a catch-up and a breakthrough. As the competition in large models enters deeper waters, those who can maintain technological sharpness while bridging the last mile of implementation will truly win the future. By entrusting the most model-savvy and product-focused young people, Baidu may be taking its boldest bet in the AGI era.