Training Object Permanence in World Models
cs.AI, cs.CV
Submitted: 2026-09-23
Updated: 2026-09-23
Comments: 26 pages, 9 figures, 5 tables. Project page: https://object-permanence.world
License: http://creativecommons.org/licenses/by/4.0/
The gist: Object permanence and solidity are hallmarks of human cognitive priors.
Terminology
Abstract
Object permanence and solidity are hallmarks of human cognitive priors. Recent studies show that video generation models, a paradigmatic class of current world models, have begun to show emerged reasoning abilities, making them ideal candidates for building human-like physical intelligence. Do video models have emerged object permanence in them? If not, could we train them with a core-cognition inspired dataset? We introduce WROP (World Reasoning with Object Permanence), a data infrastructure of 150 hand-designed cognitive science inspired tasks, divided into six cognitive categories. We build Blender generators that randomize speed, lighting, camera angle, and other nuisance parameters while preserving each task's cognitive structure, yielding 10,000+ samples per task. We release a 1.5M-sample training corpus and a 300-question exam. On this exam we evaluate 14 video models: 3 reference-to-video, 7 edit, and 4 continuation, among which PWM-WROP, our 16B world model. In a blind pairwise Elo study, PWM-WROP ranks first among continuation models and third overall, behind only a statistical tie between two reference-to-video models. We release the data, exam, model answers, scores, weights, and PWM, our native-PyTorch training stack on AWS Trainium2.
Sources
- MMGR: Multi-Modal Generative Reasoning Benchmark and Evaluation
- Are Video Models Ready as Zero-Shot Reasoners? An Empirical Study with the MME-CoF Benchmark
- LTX-2: Efficient Joint Audio-Visual Foundation Model
- RULER-Bench: Probing Rule-based Reasoning Abilities of Next-level Video Generation Models for Vision Foundation Intelligence
- HunyuanVideo: A Systematic Framework For Large Video Generative Models
- Core Knowledge Deficits in Multi-Modal Language Models
- Can World Simulators Reason? Gen-ViRe: A Generative Visual Reasoning Benchmark
- The Philosophical Foundations of Growing AI Like A Child
- Increasing Computation Resolves Conflicts in Vision Language Models
- Cosmos 3: Omnimodal World Models for Physical AI
- Cosmos World Foundation Model Platform for Physical AI
- Movie Gen: A Cast of Media Foundation Models
- Probing Perceptual Constancy in Large Vision-Language Models
- Wan: Open and Advanced Large-Scale Video Generative Models
- Video models are zero-shot learners and reasoners
- VBVR-Pro: A Scalable and Verifiable Suite for Native Visual Reasoning
- Reasoning via Video: The First Evaluation of Video Models' Reasoning Abilities through Maze-Solving Tasks
- CogVideoX: Text-to-Video Diffusion Models with An Expert Transformer
Related papers
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Agents' Last Exam
- The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
- MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
- Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems