Physics: Benchmarking Foundation Models on University-Level Physics Problem Solving
We introduce PHYSICS, a comprehensive benchmark for university-level physics problem solving.It contains 1,297 expert-annotated problems covering six core areas: classical mechanics, quantum mechanics, thermodynamics and statistical mechanics, electromagnetism, atomic physics, and optics.Each problem requires advanced physics knowledge and mathematical reasoning.We develop a robust automated evaluation system for precise and reliable validation.Our evaluation of leading foundation models reveals substantial limitations.Even the most advanced model, o3-mini, achieves only 59.9% accuracy, highlighting significant challenges in solving high-level scientific problems.Through comprehensive error analysis, exploration of diverse prompting strategies, and Retrieval-Augmented Generation (RAG)-based knowledge augmentation, we identify key areas for improvement, laying the foundation for future advancements.
Authors
- Yilun Zhao (ORCID: https://orcid.org/0000-0002-6812-5120)
- Tianyu Yang (ORCID: https://orcid.org/0000-0002-6328-8932)
- K.T. Feng (ORCID: https://orcid.org/0009-0002-4680-0604)
- Arman Cohan (ORCID: https://orcid.org/0000-0002-8954-2724)
- Chen Zhao (ORCID: https://orcid.org/0000-0003-2167-2490)
- Yixin Liu (ORCID: https://orcid.org/0000-0002-4309-5076)
- John Sous
Publication Details
- Published
- 2025-01-01
- DOI
- https://doi.org/10.18653/v1/2025.findings-acl.610
- Primary Topic
- Education and Critical Thinking Development
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- Nvidia