Beijing is advancing its strategy to become a premier global hub for scientific intelligence and artificial intelligence through a focus on open collaboration and synergistic innovation.
This initiative was highlighted during a recent thematic research activity. The city's approach involves fostering a two-way integration between AI technology and scientific research.
In recent years, Beijing has systematically developed its AI industry with the goal of establishing itself as an internationally influential center for AI innovation and industrial growth. By 2025, the number of AI enterprises in the city is projected to exceed 2,500, with the core industry scale surpassing 450 billion yuan. The city has already registered 241 large models, maintaining its leading position nationally.
The city's strategy emphasizes a dual-drive approach for both general and specialized models. This has led to significant achievements, including the release of the "Panshi" general scientific foundation model by the Chinese Academy of Sciences, the "DPA" series of atomic-scale models for the microscopic world by the Beijing Institute of Scientific Intelligence, the "MegaDFT" density functional theory model by the Zhongguancun Institute, and the "Opencomplex" series of all-atom microscopic life models by the Beijing Academy of Artificial Intelligence. The performance of these models is globally competitive, with the DPA4 model achieving top rankings on two major international benchmarks for materials science AI.
Concurrently, collaborative efforts have yielded practical platforms. The Beijing Institute of Scientific Intelligence, in partnership with Beijing Shenshi Technology Co., Ltd., has developed the "Bohr Research Space Station," the world's first end-to-end AI scientific research platform. This platform boasts over 4.5 million registered users and is deployed in more than 190 universities and research institutes, as well as over 150 enterprises. Furthermore, Shenshi Technology has launched the "Uni-Lab-OS" intelligent laboratory operating system, which supports over 1,800 types of instrument models and enables full-process experimental control via natural language.
Looking ahead, Beijing plans to leverage its resource advantages and deepen global open cooperation to build a comprehensive, closed-loop innovation ecosystem for scientific intelligence where all elements work in synergy.
A key focus will be strengthening the foundational infrastructure for scientific AI. This includes continuously upgrading scientific foundation models and strategically developing both general and specialized models. The city will also establish a number of high-level autonomous laboratories and promote the intelligent upgrading of major scientific facilities and research instruments. Accelerating the construction of top-tier scientific data centers is another priority, aiming to produce high-quality, AI-ready scientific datasets to provide comprehensive "dry-wet" closed-loop AI research support for scientists.
Beijing will also expedite the development of high-value application benchmarks. In fields such as life sciences and materials science, the city will encourage leading innovators to open their core scenarios, foster deep industry-academia-research collaboration, and organize joint research projects between enterprises, universities, research institutes, and new R&D entities. The goal is to achieve synergistic integration of scientific models, computing power, data, autonomous labs, and research agents, creating a fully closed-loop system. This effort aims to produce a batch of demonstrative, pioneering, and groundbreaking application benchmarks that will drive significant research outcomes.