PhyEditBench: A Real-World Multi-Stage Benchmark for Physics-Aware Image Editing

Researchers introduced PhyEditBench, a benchmark for evaluating physics-aware image editing in generative models, revealing significant limitations in current state-of-the-art methods. The benchmark includes 238 real-world and 35 synthetic instances, and the team proposed a training-free baseline called PhyWorld that outperforms comparable models by leveraging video generation as a reasoning mechanism.

arXiv:2606.26551v1 Announce Type: new Abstract: While instruction-based image editing, enabled by multi-modal generative models, has advanced significantly, existing benchmarks lack a comprehensive evaluation of physics-based reasoning, a critical capability for handling real-world scenarios. To address this, we introduce PhyEditBench, a benchmark designed to assess the physical understanding of editing models. Guided by a hierarchical taxonomy, we establish 4 primary classes and 12 subclasses. It comprises 238 high-quality, high-resolution, real-world instances meticulously extracted from videos to capture authentic physical dynamics, alongside 35 synthetic Anti-Physics instances. Our empirical analysis of current SOTA editing methods exposes substantial limitations in their physics-based reasoning. We further propose a training-free baseline named PhyWorld that uses test-time scaling and a latent reduction strategy. PhyWorld outperforms comparable models and suggests that the video generation process can effectively serve as a reasoning mechanism for image editing. The project page is available at https://github.com/Previsior/PhyEditBench.