Physics Supernova: AI Agent Matches Elite Gold Medalists at IPhO 2025

2025-09-01

AI Summary

Current AI systems face fundamental limitations in solving real-world physics problems requiring rigorous physical reasoning and domain-specific knowledge. Method: We propose a novel agent framework integrating physics-informed modeling, multi-step symbolic reasoning, and dynamic tool invocation. Our approach couples reinforcement learning with a symbolic engine within a prior-guided collaborative reasoning system, enabling stepwise problem modeling, analytical equation derivation, and numerical validation for complex theoretical physics problems. Results & Contributions: The agent achieves 23.5/30 on the IPhO 2025 theoretical exam—ranking 14th out of 406 globally (top 3.5%) and surpassing over half of human gold medalists. Key contributions include: (1) establishing the first IPhO-level physics reasoning benchmark; (2) empirically demonstrating substantial performance gains from principled, domain-aware tool integration in scientific reasoning; and (3) introducing an interpretable, scalable agent paradigm for AI-augmented physical discovery.

Technology Category

• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning
• Knowledge Representation and Reasoning: Reasoning with Beliefs
• Multiagent Systems: Agent/AI Theories and Architectures

Application Category

• Search and Retrieval-Augmented AI: Agentic search
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Physics provides fundamental laws that describe and predict the natural world. AI systems aspiring toward more general, real-world intelligence must therefore demonstrate strong physics problem-solving abilities: to formulate and apply physical laws for explaining and predicting physical processes. The International Physics Olympiad (IPhO)--the world's most prestigious physics competition--offers a rigorous benchmark for this purpose. We introduce Physics Supernova, an AI agent system with superior physics problem-solving abilities that match elite IPhO gold medalists. In IPhO 2025 theory problems, Physics Supernova attains 23.5/30 points, ranking 14th of 406 contestants and surpassing the median performance of human gold medalists. We extensively analyzed Physics Supernova's capabilities and flexibility across diverse physics tasks. These results show that principled tool integration within agent systems can deliver competitive improvements in solving challenging science problems. The codes are available at https://github.com/CharlesQ9/Physics-Supernova.

Problem

• Developing AI agent for physics problem-solving abilities
• Matching elite human performance in International Physics Olympiad
• Integrating tools to solve challenging science problems competitively

Innovation

• AI agent system with superior physics problem-solving
• Principled tool integration within agent systems
• Matches elite IPhO gold medalists performance