Toward Autonomous Extraction and Building of Machine-Readable Molecular Models from Publications Using Large Language Models

Abstract Force field models are one of the central pillars of molecular simulations. A large number of molecular force field models has been developed in the past decades─mostly published in scientific papers. Building machine-readable force field input files for simulation engines is a tedious and error-prone task. We developed a method for extracting and building force field files from publications using large language models (LLMs). We have tested the new method by extracting 114 force field models from 21 scientific publications. The studied force fields comprise 6–74 parameters that were automatically extracted, identified, annotated, and allocated. We have compared the performance of different LLMs, namely Gemini 2.5 Pro, Claude 4 Sonnet, Claude 3.7 Sonnet, Gemini 2.5 Flash, and GPT-4o. Overall, they yield similar performance─yet there are important differences in individual cases. The overall best performance was obtained by the Gemini 2.5 Pro LLM. The force field parameters were extracted and identified with an accuracy of 94.0% using the Gemini 2.5 Pro LLM. The new LLM-assisted extraction method drastically reduces the time required for building force field files. Nevertheless, human-in-the-loop verification is still indispensable.