PHYSICS: Benchmarking Foundation Models on University-Level Physics Problem Solving
This work addresses the need for robust evaluation of AI models in advanced physics problem-solving, though it is incremental as it applies existing benchmarking methods to a new domain.
The researchers tackled the problem of evaluating foundation models on university-level physics by creating PHYSICS, a benchmark with 1,297 expert-annotated problems across six core physics areas, and found that even the most advanced model (o3-mini) achieved only 59.9% accuracy, revealing substantial limitations in solving high-level scientific problems.
We introduce PHYSICS, a comprehensive benchmark for university-level physics problem solving. It contains 1297 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.