# IntPhys 2: Benchmarking Intuitive Physics Understanding In Complex Synthetic Environments

arXiv: https://arxiv.org/abs/2506.09849 · HTML: https://mlsanity.com/paper/2506.09849/

## Key facts

- Cited by 52 papers in the map (data through 14 Oct 2026).
- Builds on 30 of its 52 references that are in the map.
- First posted on arXiv 11 Jun 2025.

## Builds on (most cited first, 30 of 30)

- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [Qwen2.5-VL Technical Report](https://mlsanity.com/paper/2502.13923/) (2025)
- [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://mlsanity.com/paper/2104.09864/) (2021)
- [PIQA: Reasoning about Physical Commonsense in Natural Language](https://mlsanity.com/paper/1911.11641/) (2019)
- [CARLA: An Open Urban Driving Simulator](https://mlsanity.com/paper/1711.03938/) (2017)
- [HuggingFace's Transformers: State-of-the-art Natural Language Processing](https://mlsanity.com/paper/1910.03771/) (2019)
- [Datasheets for Datasets](https://mlsanity.com/paper/1803.09010/) (2018)
- [Learning Latent Dynamics for Planning from Pixels](https://mlsanity.com/paper/1811.04551/) (2018)
- [Learning to Simulate Complex Physics with Graph Networks](https://mlsanity.com/paper/2002.09405/) (2020)
- [AI2-THOR: An Interactive 3D Environment for Visual AI](https://mlsanity.com/paper/1712.05474/) (2017)
- [Cosmos World Foundation Model Platform for Physical AI](https://mlsanity.com/paper/2501.03575/) (2025)
- [Revisiting Feature Prediction for Learning Visual Representations from Video](https://mlsanity.com/paper/2404.08471/) (2024)
- [VideoMAE V2: Scaling Video Masked Autoencoders with Dual Masking](https://mlsanity.com/paper/2303.16727/) (2023)
- [CLEVRER: CoLlision Events for Video REpresentation and Reasoning](https://mlsanity.com/paper/1910.01442/) (2019)
- [VideoPhy: Evaluating Physical Commonsense for Video Generation](https://mlsanity.com/paper/2406.03520/) (2024)
- [An Introduction to Vision-Language Modeling](https://mlsanity.com/paper/2405.17247/) (2024)
- [Do generative video models understand physical principles?](https://mlsanity.com/paper/2501.09038/) (2025)
- [Physion: Evaluating Physical Prediction from Vision in Humans and Machines](https://mlsanity.com/paper/2106.08261/) (2021)
- [PHYRE: A New Benchmark for Physical Reasoning](https://mlsanity.com/paper/1908.05656/) (2019)
- [CATER: A diagnostic dataset for Compositional Actions and TEmporal Reasoning](https://mlsanity.com/paper/1910.04744/) (2019)
- [Oops! Predicting Unintentional Action in Video](https://mlsanity.com/paper/1911.11206/) (2019)
- [Intuitive physics understanding emerges from self-supervised pretraining on natural videos](https://mlsanity.com/paper/2502.11831/) (2025)
- [Visual Grounding of Learned Physical Models](https://mlsanity.com/paper/2004.13664/) (2020)
- [Physion++: Evaluating Physical Scene Understanding that Requires Online Inference of Different Physical Properties](https://mlsanity.com/paper/2306.15668/) (2023)
- [PUG: Photorealistic and Semantically Controllable Synthetic Data for Representation Learning](https://mlsanity.com/paper/2308.03977/) (2023)
- [GRASP: A novel benchmark for evaluating language GRounding And Situated Physics understanding in multimodal language models](https://mlsanity.com/paper/2311.09048/) (2023)
- [LLMs Are Biased Towards Output Formats! Systematically Evaluating and Mitigating Output Format Bias of LLMs](https://mlsanity.com/paper/2408.08656/) (2024)
- [WorldSense: A Synthetic Benchmark for Grounded Reasoning in Large Language Models](https://mlsanity.com/paper/2311.15930/) (2023)
- [Compositional 4D Dynamic Scenes Understanding with Physics Priors for Video Question Answering](https://mlsanity.com/paper/2406.00622/) (2024)
- [Occlusion resistant learning of intuitive physics from videos](https://mlsanity.com/paper/2005.00069/) (2020)

## Built on by (most cited first, 30 of 52)

- [World Simulation with Video Foundation Models for Physical AI](https://mlsanity.com/paper/2511.00062/) (2025)
- [Embodied AI Agents: Modeling the World](https://mlsanity.com/paper/2506.22355/) (2025)
- [Inference-time Physics Alignment of Video Generative Models with Latent World Models](https://mlsanity.com/paper/2601.10553/) (2026)
- [PAI-Bench: A Comprehensive Benchmark For Physical AI](https://mlsanity.com/paper/2512.01989/) (2025)
- [Generative Universal Verifier as Multimodal Meta-Reasoner](https://mlsanity.com/paper/2510.13804/) (2025)
- [LikePhys: Evaluating Intuitive Physics Understanding in Video Diffusion Models via Likelihood Preference](https://mlsanity.com/paper/2510.11512/) (2025)
- [Code as Worlds: Agentic Discovery of Executable World Representations for Physical Reasoning](https://mlsanity.com/paper/2608.27549/) (2026)
- [Rethinking JEPA: Compute-Efficient Video SSL with Frozen Teachers](https://mlsanity.com/paper/2509.24317/) (2025)
- [WorldBench: Benchmarking Physical Understanding of World Models by Isolating Physics Concepts](https://mlsanity.com/paper/2601.21282/) (2026)
- [PhysInOne: Visual Physics Learning and Reasoning in One Suite](https://mlsanity.com/paper/2604.09415/) (2026)
- [Mind Meets Space: Rethinking Agentic Spatial Intelligence from a Neuroscience-inspired Perspective](https://mlsanity.com/paper/2509.09154/) (2025)
- [VisPhyWorld: Probing Physical Reasoning via Code-Driven Video Reconstruction](https://mlsanity.com/paper/2602.13294/) (2026)
- [Inferring Dynamic Physical Properties from Video Foundation Models](https://mlsanity.com/paper/2510.02311/) (2025)
- [Learning Physics-Grounded 4D Dynamics with Neural Gaussian Force Fields](https://mlsanity.com/paper/2602.00148/) (2026)
- [VKnowU: Evaluating Visual Knowledge Understanding in Multimodal LLMs](https://mlsanity.com/paper/2511.20272/) (2025)
- [TRAPSBench: Vision-Language Models Encode but Fail to Express Epistemic Restraint](https://mlsanity.com/paper/2608.13167/) (2026)
- [CRONOS: Benchmarking Counterfactual Physical Consistency in Video Models](https://mlsanity.com/paper/2605.23699/) (2026)
- [PhysMind: From Video to Executable Worlds for Training-Free Physical Reasoning](https://mlsanity.com/paper/2608.04575/) (2026)
- [Pixels to Principles: Probing Intuitive Physics Understanding in Multimodal Language Models](https://mlsanity.com/paper/2507.16572/) (2025)
- [Next-Gen CAPTCHAs: Leveraging the Cognitive Gap for Scalable and Diverse GUI-Agent Defense](https://mlsanity.com/paper/2602.09012/) (2026)
- [YoCausal: How Far is Video Generation from World Model? A Causality Perspective](https://mlsanity.com/paper/2605.30346/) (2026)
- [GEOPHYS: The Geometry of Physical Plausibility](https://mlsanity.com/paper/2606.20707/) (2026)
- [4DCodeBench: Benchmarking Agents on Inverse Graphics of Dynamic Scenes](https://mlsanity.com/paper/2610.03715/) (2026)
- [InPhyRe Discovers: Large Multimodal Models Struggle in Inductive Physical Reasoning](https://mlsanity.com/paper/2509.12263/) (2025)
- [PhysEditWorld: A Large-Scale Dataset Toward Physics-Editable World Models](https://mlsanity.com/paper/2606.26694/) (2026)
- [SiPhy: Single-Image Physical Property Reasoning](https://mlsanity.com/paper/2607.22355/) (2026)
- [PhiZero: A World Model Built Around Physical Language](https://mlsanity.com/paper/2607.28624/) (2026)
- [Principia: Relational Physics Tests for Video Models](https://mlsanity.com/paper/2609.04200/) (2026)
- [PhysFire-WM: A Physics-Informed World Model for Emulating Fire Spread Dynamics](https://mlsanity.com/paper/2512.17152/) (2025)
- [Modeling the Mental World for Embodied AI: A Comprehensive Review](https://mlsanity.com/paper/2601.02378/) (2025)

## Abstract (arXiv)

We present IntPhys 2, a video benchmark designed to evaluate the intuitive physics understanding of deep learning models. Building on the original IntPhys benchmark, IntPhys 2 focuses on four core principles related to macroscopic objects: Permanence, Immutability, Spatio-Temporal Continuity, and Solidity. These conditions are inspired by research into intuitive physical understanding emerging during early childhood. IntPhys 2 offers a comprehensive suite of tests, based on the violation of expectation framework, that challenge models to differentiate between possible and impossible events within controlled and diverse virtual environments. Alongside the benchmark, we provide performance evaluations of several state-of-the-art models. Our findings indicate that while these models demonstrate basic visual understanding, they face significant challenges in grasping intuitive physics across the four principles in complex scenes, with most models performing at chance levels (50%), in stark contrast to human performance, which achieves near-perfect accuracy. This underscores the gap between current models and human-like intuitive physics understanding, highlighting the need for advancements in model architectures and training methodologies.