On April 28, the Chinese Academy of Sciences officially launched the "Panzhi 100" model system in Beijing. This system is built upon the "Panzhi" scientific foundational model and encompasses a cluster of discipline-specific large models covering eight key fields: mathematics, physics, materials science, astronomy, environmental science, aerospace, geography, and life sciences. This initiative marks the official transition of AI for Science from previous fragmented and closed single-point exploration models to a collaborative and efficient platform-based innovation stage.

Core Base Upgrade, Three Functions to Support Innovation

As the intelligent base of the entire system, the "Panzhi Scientific Foundation Model" has been upgraded to version 1.5pro. Based on 6.5 million high-quality scientific reasoning data, the model's logical reasoning, multimodal understanding, and result reliability have been significantly improved. The platform currently provides three core functions: "Literature Compass," "Innovation Evaluation," and "Intelligent Factory," aiming to streamline the entire research innovation process. Among them, the "Intelligent Factory" has accumulated over 2,000 research tools, achieving an initial self-sufficient tool chain and offering researchers a one-stop intelligent assistance solution.

Eight Disciplinary Models Each Show Their Strength, Accelerating Breakthroughs

In specific disciplinary applications, "Panzhi 100" has demonstrated strong professional penetration. For example, the "Dayan Zhizheng" model in mathematics can support complex fluid simulation calculations; the "Doctor Sai" model in physics delves into particle physics analysis and is expected to redefine the research paradigms of large scientific facilities. In materials science, the "Zhurong" model has achieved on-demand design and precise preparation of new materials; the "Jinyu" model in astronomy can automatically predict solar flares.

Additionally, the "Yuheng" system in environmental science has achieved the world's first panoramic carbon emission accounting; the "Lingkong" model in aerospace has deep cognitive and complex reasoning capabilities; the "Kunyuan" model in geography has helped develop the first national standard for landform classification coding. In life sciences, the "Digital Cell" model has shown remarkable performance, discovering three unknown drug targets within 30 days and successfully verifying them through experiments, providing a new path for personalized and precise diagnosis and treatment.

Interdisciplinary Collaboration, Building a New Digital Research Ecosystem