# PhyGround: Benchmarking Physical Reasoning in Generative World Models

arXiv: https://arxiv.org/abs/2605.10806 · HTML: https://mlsanity.com/paper/2605.10806/

## Key facts

- Cited by 12 papers in the map (data through 14 Oct 2026).
- Placed in Latent predictive world models, one of 11 directions in World models.
- Builds on 27 of its 51 references that are in the map.
- First posted on arXiv 11 May 2026.

## Placed in

- World models / [Latent predictive world models](https://mlsanity.com/wm/latent-predictive-world-models/)

## Builds on (most cited first, 27 of 27)

- [Qwen3-VL Technical Report](https://mlsanity.com/paper/2511.21631/) (2025)
- [Wan: Open and Advanced Large-Scale Video Generative Models](https://mlsanity.com/paper/2503.20314/) (2025)
- [Gemini: A Family of Highly Capable Multimodal Models](https://mlsanity.com/paper/2312.11805/) (2023)
- [VBench-2.0: Advancing Video Generation Benchmark Suite for Intrinsic Faithfulness](https://mlsanity.com/paper/2503.21755/) (2025)
- [Improving Video Generation with Human Feedback](https://mlsanity.com/paper/2501.13918/) (2025)
- [LTX-2: Efficient Joint Audio-Visual Foundation Model](https://mlsanity.com/paper/2601.03233/) (2026)
- [VideoPhy: Evaluating Physical Commonsense for Video Generation](https://mlsanity.com/paper/2406.03520/) (2024)
- [Towards World Simulator: Crafting Physical Commonsense-Based Benchmark for Video Generation](https://mlsanity.com/paper/2410.05363/) (2024)
- [VideoPhy-2: A Challenging Action-Centric Physical Commonsense Evaluation in Video Generation](https://mlsanity.com/paper/2503.06800/) (2025)
- [WorldModelBench: Judging Video Generation Models As World Models](https://mlsanity.com/paper/2502.20694/) (2025)
- [Evaluating Newtonian Mechanics in Video Generative Models with Real Physical Systems](https://mlsanity.com/paper/2504.02918/) (2025)
- [VideoScore2: Think before You Score in Generative Video Evaluation](https://mlsanity.com/paper/2509.22799/) (2025)
- [T2VPhysBench: A First-Principles Benchmark for Physical Consistency in Text-to-Video Generation](https://mlsanity.com/paper/2505.00337/) (2025)
- ["PhyWorldBench": A Comprehensive Evaluation of Physical Realism in Text-to-Video Models](https://mlsanity.com/paper/2507.13428/) (2025)
- [A Very Big Video Reasoning Suite](https://mlsanity.com/paper/2602.20159/) (2026)
- [OmniWeaving: Towards Unified Video Generation with Free-form Composition and Reasoning](https://mlsanity.com/paper/2603.24458/) (2026)
- [PhysGame: Uncovering Physical Commonsense Violations in Gameplay Videos](https://mlsanity.com/paper/2412.01800/) (2024)
- [TRAVL: A Recipe for Making Video-Language Models Better Judges of Physics Implausibility](https://mlsanity.com/paper/2510.07550/) (2025)
- [MJ-VIDEO: Fine-Grained Benchmarking and Rewarding Video Preferences in Video Generation](https://mlsanity.com/paper/2502.01719/) (2025)
- [Impossible Videos](https://mlsanity.com/paper/2503.14378/) (2025)
- [Physion-Eval: Evaluating Physical Realism in Generated Video via Human Reasoning](https://mlsanity.com/paper/2603.19607/) (2026)
- [Content-Rich AIGC Video Quality Assessment via Intricate Text Alignment and Motion-Aware Consistency](https://mlsanity.com/paper/2502.04076/) (2025)
- [LOVE: Benchmarking and Evaluating Text-to-Video Generation and Video-to-Text Interpretation](https://mlsanity.com/paper/2505.12098/) (2025)
- [Benchmarking Scientific Understanding and Reasoning for Video Generation using VideoScience-Bench](https://mlsanity.com/paper/2512.02942/) (2025)
- [PhyDetEx: Detecting and Explaining the Physical Plausibility of T2V Models](https://mlsanity.com/paper/2512.01843/) (2025)
- [Human Cognition in Machines: A Unified Perspective of World Models](https://mlsanity.com/paper/2604.16592/) (2026)
- [AIGVE-MACS: Unified Multi-Aspect Commenting and Scoring Model for AI-Generated Video Evaluation](https://mlsanity.com/paper/2507.01255/) (2025)

## Built on by (most cited first, 12 of 12)

- [WorldExam: Benchmarking World Models from Apparent Appearance to Inherent Reactivity](https://mlsanity.com/paper/2608.02603/) (2026)
- [RigidBench: Evaluating Rigid-Body Physics in Video Generation Models](https://mlsanity.com/paper/2608.15555/) (2026)
- [PhiZero: A World Model Built Around Physical Language](https://mlsanity.com/paper/2607.28624/) (2026)
- [PhysWeep: Does a Video Generator Realize the Physics You Ask For?](https://mlsanity.com/paper/2609.06207/) (2026)
- [Foresight at the Event Boundary: Evaluating Physical Prediction in Video World Models](https://mlsanity.com/paper/2609.36531/) (2026)
- [SCOPE: Score-Isolated Agentic Optimization for Video World Models](https://mlsanity.com/paper/2608.15043/) (2026)
- [VeriPhy: Agentic Physical Reasoning for World Model Evaluation and Refinement](https://mlsanity.com/paper/2609.03153/) (2026)
- [Can MiniMax-H3 Reason About the Physical World? An Evaluation of Omni-Modal Generative Model](https://mlsanity.com/paper/2609.18323/) (2026)
- [Conservation Buys Stability and Factoring Buys Counterfactuals in Physical World Models](https://mlsanity.com/paper/2609.19674/) (2026)
- [Physis-Lang: Self-Evolving Language as a Physical Representation for Video World Model](https://mlsanity.com/paper/2609.40358/) (2026)
- [S2PD: Serial-to-Parallel Diffusion for Physically and Logically Consistent Video Generation](https://mlsanity.com/paper/2610.06847/) (2026)
- [World Models' Last Exam in Physics](https://mlsanity.com/paper/2610.08791/) (2026)

## Abstract (arXiv)

Generative world models are increasingly used for video generation, where learned simulators are expected to capture the physical rules that govern real-world dynamics. However, evaluating whether generated videos actually follow these rules remains challenging. Existing physics-focused video benchmarks have made important progress, but they still face three key challenges, including the coarse evaluation frameworks that hide law-specific failures, response biases and fatigue that undermine the validity of annotation judgments, and automated evaluators that are insufficiently physics-aware or difficult to audit. To address those challenges, we introduce PhyGround, a criteria-grounded benchmark for evaluating physical reasoning in video generation. The benchmark contains 250 curated prompts, each augmented with an expected physical outcome, and a taxonomy of 13 physical laws across solid-body mechanics, fluid dynamics, and optics. Each law is operationalized through observable sub-questions to enable per-law diagnostics. We evaluate eight modern video generation models through a large-scale, quality-controlled human study, grounded on social science lab experiment design. A total of 459 annotators provided 5,796 complete annotations and over 37.4K fine-grained labels; after quality control, the retained annotations exhibited high split-half model-ranking correlations (Spearman's rho > 0.90). To support reproducible automated evaluation, we release PhyJudge-9B, an open physics-specialized VLM judge. PhyJudge-9B achieves substantially lower aggregate relative bias than Gemini-3.1-Pro (3.3% vs. 16.6%). We release prompts, human annotations, model checkpoints, and evaluation code on the project page https://phyground.github.io/.