MatOS opens source for AI-assisted materials research
Beijing Zhongguancun Academy and the Zhongguancun Institute of Artificial Intelligence have open-sourced MatOS, an agent system for materials research. Its asset library contains 3,135 datasets, 386 skills, 104 workflows and 51 expert roles, while the system integrates 48 materials-computing applications. The project team reports that MatOS led compared systems on ResearchClawBench and MatTools, but the supplied reports provide no independent replication.
MatOS combines reusable scientific assets, stateful workflows and evidence-driven asset revision in an open materials-research agent system, alongside published code, a technical report and integrations with 48 materials-computing tools.
大模型智能体已经能够检索文献、调用模拟软件,甚至与自动化实验室协同工作。但要让它们独立完成一项完整的材料研究,仍然面临挑战:
一项研究可能要从数千万个候选结构中逐级筛选,或在数百个合金配方中反复搜索,持续数天,并根据中间结果不断调整计划;每接入一个新的数据库或软件,往往要重新搭建流程;一项研究中积累的经验,也很难被其他研究复用。
为此,北京中关村学院、中关村人工智能研究院(中关村两院)于今年九月发布年度科研进展——MatOS,一个面向材料研发的科研智能体系统。
该项目由中关村两院AI材料研究所团队构建,将数据、知识、工具、技能、角色和工作流统一组织为科学资产,支持研究者和MatOS自身在已有资产之上构建新能力,通过有状态工作流执行研究任务,并将经过检查与审查的资产修订用于后续研究,使科研能力在使用中持续进化。
#MatOS#Beijing Zhongguancun Academy#Zhongguancun Institute of Artificial Intelligence#AI Materials Institute#DuoCompNet#ResearchClawBench#MatTools#SciAgentGym#RarityGuard-MP#GPT-5.4#DeepSeek-V4-Flash#DeepSeek-V4-Pro