WorldModelBench: Judging Video Generation Models As World Models
This addresses the need for better evaluation of video generation models as world models for applications like robotics and autonomous driving, though it is incremental as it builds on existing benchmarking efforts.
The authors tackled the problem that existing benchmarks for video generation models fail to rigorously evaluate their world modeling capabilities, such as physics adherence, and proposed WorldModelBench, a new benchmark that incorporates instruction-following and physics-adherence dimensions, achieving 8.6% higher average accuracy in predicting violations than GPT-4o with 2B parameters.
Video generation models have rapidly progressed, positioning themselves as video world models capable of supporting decision-making applications like robotics and autonomous driving. However, current benchmarks fail to rigorously evaluate these claims, focusing only on general video quality, ignoring important factors to world models such as physics adherence. To bridge this gap, we propose WorldModelBench, a benchmark designed to evaluate the world modeling capabilities of video generation models in application-driven domains. WorldModelBench offers two key advantages: (1) Against to nuanced world modeling violations: By incorporating instruction-following and physics-adherence dimensions, WorldModelBench detects subtle violations, such as irregular changes in object size that breach the mass conservation law - issues overlooked by prior benchmarks. (2) Aligned with large-scale human preferences: We crowd-source 67K human labels to accurately measure 14 frontier models. Using our high-quality human labels, we further fine-tune an accurate judger to automate the evaluation procedure, achieving 8.6% higher average accuracy in predicting world modeling violations than GPT-4o with 2B parameters. In addition, we demonstrate that training to align human annotations by maximizing the rewards from the judger noticeably improve the world modeling capability. The website is available at https://worldmodelbench-team.github.io.