T2VPhysBench: A First-Principles Benchmark for Physical Consistency in Text-to-Video Generation

TL;DR

This paper introduces T2VPhysBench, a comprehensive benchmark for evaluating whether text-to-video models adhere to fundamental physical laws, revealing widespread violations and guiding future improvements.

Contribution

It presents the first systematic, physics-based evaluation benchmark for text-to-video models, combining human judgment and rigorous testing of core physical principles.

Findings

All models scored below 0.60 on average in physical law compliance.

Law-specific hints do not significantly improve physical adherence.

Models often violate physical laws when explicitly instructed to do so.

Abstract

Text-to-video generative models have made significant strides in recent years, producing high-quality videos that excel in both aesthetic appeal and accurate instruction following, and have become central to digital art creation and user engagement online. Yet, despite these advancements, their ability to respect fundamental physical laws remains largely untested: many outputs still violate basic constraints such as rigid-body collisions, energy conservation, and gravitational dynamics, resulting in unrealistic or even misleading content. Existing physical-evaluation benchmarks typically rely on automatic, pixel-level metrics applied to simplistic, life-scenario prompts, and thus overlook both human judgment and first-principles physics. To fill this gap, we introduce \textbf{T2VPhysBench}, a first-principled benchmark that systematically evaluates whether state-of-the-art text-to-video systems, both open-source and commercial, obey twelve core physical laws including Newtonian mechanics, conservation principles, and phenomenological effects. Our benchmark employs a rigorous human evaluation protocol and includes three targeted studies: (1) an overall compliance assessment showing that all models score below 0.60 on average in each law category; (2) a prompt-hint ablation revealing that even detailed, law-specific hints fail to remedy physics violations; and (3) a counterfactual robustness test demonstrating that models often generate videos that explicitly break physical rules when so instructed. The results expose persistent limitations in current architectures and offer concrete insights for guiding future research toward truly physics-aware video generation.