PhysCodeBench: Benchmarking Physics-Aware Symbolic Simulation of 3D Scenes via Self-Corrective Multi-Agent Refinement

2026-04-26

AI Summary

This work addresses the semantic gap between natural language descriptions and executable 3D physical simulations by proposing a Self-correcting Multi-agent Refinement Framework (SMRF). SMRF employs three specialized agents—simulation generation, error correction, and refinement—that iteratively collaborate under a domain-specific validation mechanism to produce high-fidelity physics code. The study introduces PhysCodeBench, the first comprehensive benchmark comprising 700 samples spanning rigid-body dynamics, fluid mechanics, and soft-body physics, alongside a novel paradigm for physics simulation generation based on multi-agent collaboration and self-correction. Experimental results demonstrate that SMRF achieves a score of 67.7 on PhysCodeBench, outperforming the strongest baseline by 31.4 points, thereby substantiating the framework’s critical role in enhancing the accuracy of physically grounded symbolic simulations.

Technology Category

• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Multiagent Systems: Agent-Based Simulation and Emergent Behavior
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Search and Retrieval-Augmented AI: Agentic search
• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Physics-aware symbolic simulation of 3D scenes is critical for robotics, embodied AI, and scientific computing, requiring models to understand natural language descriptions of physical phenomena and translate them into executable simulation environments. While large language models (LLMs) excel at general code generation, they struggle with the semantic gap between physical descriptions and simulation implementation. We introduce PhysCodeBench, the first comprehensive benchmark for evaluating physics-aware symbolic simulation, comprising 700 manually-crafted diverse samples across mechanics, fluid dynamics, and soft-body physics with expert annotations. Our evaluation framework measures both code executability and physical accuracy through automated and visual assessment. Building on this, we propose a Self-Corrective Multi-Agent Refinement Framework (SMRF) with three specialized agents (simulation generator, error corrector, and simulation refiner) that collaborate iteratively with domain-specific validation to produce physically accurate simulations. SMRF achieves 67.7 points overall performance compared to 36.3 points for the best baseline among evaluated SOTA models, representing a 31.4-point improvement. Our analysis demonstrates that error correction is critical for accurate physics-aware symbolic simulation and that specialized multi-agent approaches significantly outperform single-agent methods across the tested physical domains.

Problem

• physics-aware symbolic simulation
• 3D scene simulation
• semantic gap
• physical accuracy
• executable code generation

Innovation

• physics-aware simulation
• symbolic code generation
• multi-agent refinement
• self-corrective framework
• 3D scene simulation