# Benchmarking World-Model Learning with Environment-Level Queries

arXiv: https://arxiv.org/abs/2510.19788 · HTML: https://mlsanity.com/paper/2510.19788/

## Key facts

- Cited by 10 papers in the map (data through 14 Oct 2026).
- Placed in World models for language agents, one of 11 directions in World models.
- Builds on 22 of its 51 references that are in the map.
- First posted on arXiv 22 Oct 2025.

## Placed in

- World models / [World models for language agents](https://mlsanity.com/wm/world-models-for-language-agents/)

## Builds on (most cited first, 22 of 22)

- [SWE-bench: Can Language Models Resolve Real-World GitHub Issues?](https://mlsanity.com/paper/2310.06770/) (2023)
- [GPQA: A Graduate-Level Google-Proof Q&A Benchmark](https://mlsanity.com/paper/2311.12022/) (2023)
- [VBench: Comprehensive Benchmark Suite for Video Generative Models](https://mlsanity.com/paper/2311.17982/) (2023)
- [OpenAI Gym](https://mlsanity.com/paper/1606.01540/) (2016)
- [Evaluating Real-World Robot Manipulation Policies in Simulation](https://mlsanity.com/paper/2405.05941/) (2024)
- [Unsupervised Learning of Video Representations using LSTMs](https://mlsanity.com/paper/1502.04681/) (2015)
- [dm_control: Software and Tasks for Continuous Control](https://mlsanity.com/paper/2006.12983/) (2020)
- [RAVEN: A Dataset for Relational and Analogical Visual rEasoNing](https://mlsanity.com/paper/1903.02741/) (2019)
- [In-context Reinforcement Learning with Algorithm Distillation](https://mlsanity.com/paper/2210.14215/) (2022)
- [Measuring abstract reasoning in neural networks](https://mlsanity.com/paper/1807.04225/) (2018)
- [Self-Supervised Visual Planning with Temporal Skip Connections](https://mlsanity.com/paper/1710.05268/) (2017)
- [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)
- [DISCOVERYWORLD: A Virtual Environment for Developing and Evaluating Automated Scientific Discovery Agents](https://mlsanity.com/paper/2406.06769/) (2024)
- [MiniHack the Planet: A Sandbox for Open-Ended Reinforcement Learning Research](https://mlsanity.com/paper/2109.13202/) (2021)
- [Alchemy: A benchmark and analysis toolkit for meta-reinforcement learning agents](https://mlsanity.com/paper/2102.02926/) (2021)
- [Solving Bongard Problems with a Visual Language and Pragmatic Reasoning](https://mlsanity.com/paper/1804.04452/) (2018)
- [Assessing Adaptive World Models in Machines with Novel Games](https://mlsanity.com/paper/2507.12821/) (2025)
- [PUZZLES: A Benchmark for Neural Algorithmic Reasoning](https://mlsanity.com/paper/2407.00401/) (2024)
- [Imagine the Unseen World: A Benchmark for Systematic Generalization in Visual World Models](https://mlsanity.com/paper/2311.09064/) (2023)
- [Context and Diversity Matter: The Emergence of In-Context Learning in World Models](https://mlsanity.com/paper/2509.22353/) (2025)
- [Software Engineering User Study Recruitment on Prolific: An Experience Report](https://mlsanity.com/paper/2201.05348/) (2022)

## Built on by (most cited first, 10 of 10)

- [AI Gamestore: Scalable, Open-Ended Evaluation of Machine General Intelligence with Human Games](https://mlsanity.com/paper/2602.17594/) (2026)
- [PhysicsMind: Sim and Real Mechanics Benchmarking for Physical Reasoning and Prediction in Foundational VLMs and World Models](https://mlsanity.com/paper/2601.16007/) (2026)
- [Medical world models in healthcare: foundations, applications, and challenges for trustworthy clinical translation](https://mlsanity.com/paper/2607.25242/) (2026)
- [MaD Physics: Evaluating information seeking under constraints in physical environments](https://mlsanity.com/paper/2605.10820/) (2026)
- [AutoWorldModel-Bench: A State-Centric Benchmark for Automated World-Model Research](https://mlsanity.com/paper/2608.11216/) (2026)
- [Multi-Horizon Consistency as Geometry: When Latent Dynamics Contract, and When They Do Not](https://mlsanity.com/paper/2607.21645/) (2026)
- [Schema: Discovering Unknown Environments via Agentic Program Induction](https://mlsanity.com/paper/2609.39140/) (2026)
- [Open-Endedness Bench: Measuring Epistemic Process from Agent Records](https://mlsanity.com/paper/2610.02588/) (2026)
- [ULTRADISCOVERY: Abductive Exploration in an Interconnected, Epistemically Open Universe](https://mlsanity.com/paper/2610.03092/) (2026)
- [WOVEN: Weaving Visual World Modeling into Multimodal LLMs](https://mlsanity.com/paper/2610.12417/) (2026)

## Abstract (arXiv)

World models are central to building AI agents capable of flexible reasoning and planning. Yet current evaluations (i) test only properties measurable from observed interactions, such as next-frame prediction or task return, and (ii) do not test whether a learned model supports diverse queries about the environment. In contrast, humans build textitgeneral-purpose models that can answer many different questions about an environmentunicodex2014including questions that require understanding global structure and counterfactual consequences. We propose textitWorldTest: a protocol for evaluating whether agents learn models that support multiple textitenvironment-level queriesunicodex2014questions whose answers depend on properties of the full environment, not just observed trajectories. Individually, these queries can target properties (e.g., reachability or the effects of interventions) that no single rollout distribution determines. Collectively, they assess model generality across query types. We instantiate WorldTest as textitAutumnBench, a benchmark of 43 interactive grid-world environments and 129 tasks across three query families for both humans and learning agents. Experiments with 517 human participants and five frontier models show that humans substantially outperform these models, a gap we attribute to differences in exploration and belief updating. AutumnBench provides a framework for evaluating world-model learning in grid-world environments with environment-level queries, and WorldTest provides a template for extending such evaluations to richer domains.