THE latest version of China’s first large language model dedicated to the chemical industry was recently released, marking a significant evolution from knowledge acquisition to intelligent execution of complex chemical tasks.

The ChemELLM 3.0 Pro, jointly developed by the Dalian Institute of Chemical Physics (DICP) under the Chinese Academy of Sciences, iFlytek, Alibaba Cloud Computing, and other institutions, now functions as an intelligent engineering partner that can collaborate with professionals on intricate operations.

The model has already gained substantial traction, with more than 300 chemical enterprises, universities, and research institutes registering to use it, and cumulative API calls exceeding 14 million to date. The chemical industry involves complex materials, large spans in scale, diverse data types, and stringent engineering constraints, making it difficult for general-purpose large models to meet professional needs.

To address these challenges, the new version breaks away from the conventional application model centered on knowledge Q&A and content generation. It establishes a four-tier architecture comprising a large model, intelligent agents, professional skills and tools, and application scenarios. In this architecture, the large model functions as the cognitive core, responsible for professional knowledge comprehension, multimodal information parsing, solution generation, and task planning.

Intelligent agents act as execution entities, handling process decomposition, tool invocation, result verification, and dynamic adjustments around task objectives, forming a closed loop that integrates cognition, reasoning, planning, execution, and validation.

According to assessments by a chemical-domain evaluation system, ChemELLM 3.0 Pro achieved text-based Q&A accuracy of 81.96% and multimodal Q&A accuracy of 80.75%, with overall scores showing respective improvements of 20.2% and 31.4% compared with the 3.0 version. (SD-Agencies)