mlsanityData to 7 Oct 2026

World models for language agents

486 papers3 this weekLow confidence in this grouping: the papers may belong to more than one topic, or the group may changeLow confidence in this grouping: the papers may belong to more than one topic, or the group may change · About

Keywords: reasoning, agents, language, llms, agent

Key facts. World models for language agents: 486 papers on arXiv since 2013 (data through 7 Oct 2026). 81 of its 486 mapped papers (17%) cite “Reasoning with Language Model is Planning with World Model” (2023, arXiv 2305.14992). 74 of its 486 mapped papers (15%) cite “Language Models are Few-Shot Learners” (2020, arXiv 2005.14165). 35 new papers 17 Sep–7 Oct.

What’s new

17 Sep – 7 Oct 2026

35 new papers, 17 Sep – 7 Oct; 3 this week.

  1. This week · 3
  2. MemoWM: How World Models Change What Agents Need to Remember
  3. Long-Horizon Textual World Modeling through Structured Reasoning
  4. From Traces to Agentic Worlds: Agentic Language World Models for Interactive Environment Simulation
  5. Earlier · 32
  6. When Low Prediction Error Misleads Planning: Diagnosing Representation, Dynamics, and Decision Failures in Latent World Models also fits: Latent predictive world models
  7. What Should World Models Forget? Stratified Retention for Continual Adaptation
  8. ReFract: Benchmarking Perspective Awareness in Language Model Agents with Text World Models
  9. EpiWorld: Grounding LLM Policy Agents in Epidemiological World Models
  10. CuBEs: Culturally-Situated Behavioral Evaluations and the Limitations of Culture-Blind LLM Judges
  11. How To Train Your World Model: Fine-tuning vs RAG for LM-based World Modeling
  12. World Editing: Intervening on Executable Worlds at Increasing Depth also fits: Video-generation world models and simulators
All 35
  1. On the Divergence of Accuracy and Mechanism Consistency in Time Series World Models also fits: Latent predictive world models
  2. In CEM, a World Model Is Also a Proposal Mechanism also fits: Latent dynamics models and model-based RL
  3. EVOKE: Eliciting World Knowledge in Agents for Transferable Decision-Making
  4. Kepler: Auditable World Models for ARC-AGI-3 also fits: World model security and GUI agents
  5. Asking the World: Generalist Physical Reasoning through Agentic World Modeling and Probing also fits: Latent predictive world models
  6. Not All Errors Matter: Decision-Relevant Prediction Error Predicts Planning Quality also fits: Latent dynamics models and model-based RL
  7. Lucid Dreaming for World Models: Learning to Doubt Imagination and Decide by Trust
  8. RolloutFaith: Auditing Persistent Internal Interventions in Visual World Model also fits: Latent predictive world models
  9. Graph World Models for Constrained Epidemic Policy Planning also fits: Latent dynamics models and model-based RL
  10. RoboFL: Federated Expert Assembly for World Action Models also fits: Embodied and robot world models
  11. WM-VLM: Probing Internal World Models for Interleaved Visual-Textual Reasoning also fits: World models for navigation
  12. Shaping Persistent Representations from Independent Interactions also fits: Latent predictive world models
  13. WorldGraph: Graph-Native World Modeling also fits: Latent predictive world models
  14. Do World Models Learn Global Understanding?
  15. ALDER: Discovering the Laws of a World by Acting in It
  16. What Must a World Model Distinguish for Planning? also fits: Latent dynamics models and model-based RL
  17. World Agent: Can Language Models Keep a World Running?
  18. Seeing Parts, Reasoning about Worlds: Visual Inference under Partial Observation also fits: Video-generation world models and simulators
  19. COUNTERMEM: World-Model Verified Counter-Factual Memory for Language Agents
  20. Training Object Permanence in World Models also fits: Latent predictive world models
  21. Agent-Editing World Model: Rethinking World Modeling for LLM Agents
  22. Dual-Frontier: When Can an Agent Trust Its World Model? also fits: Latent dynamics models and model-based RL
  23. Beyond Visual Quality: A Study of Test-Time Planning with World Action Models also fits: World models for navigation
  24. World Modeling in Transformers also fits: World models for navigation
  25. Continual Enterprise World Model Discovery in Dynamic Systems

Foundations

Most cited by papers here

  1. Reasoning with Language Model is Planning with World Model2023cited by 81 here
  2. Language Models are Few-Shot Learners2020cited by 74 here
  3. Generative Agents: Interactive Simulacra of Human Behavior2023cited by 52 here
  4. ALFWorld: Aligning Text and Embodied Environments for Interactive Learning2020cited by 45 here
  5. Understanding World or Predicting Future? A Comprehensive Survey of World Models2024cited by 36 here

Lineage

The papers this direction builds on, oldest first.

Each paper on the line cites the one before.

  1. 2018World Modelscited by 148 here
  2. 2018Learning Latent Dynamics for Planning from Pixelscited by 75 here
  3. 2019Dream to Control: Learning Behaviors by Latent Imaginationcited by 110 here
  4. 2020Mastering Atari with Discrete World Modelscited by 69 here
  5. 2022Transformers are Sample-Efficient World Modelscited by 39 here
  6. 2023Mastering Diverse Domains through World Modelscited by 93 here
  7. 2024Web Agents with World Models: Learning and Leveraging Environment Dynamics in Web Navigationcited by 32 here
  8. 2024Is Your LLM Secretly a World Model of the Internet? Model-Based Planning for Web Agentscited by 32 here
  9. 2024Understanding World or Predicting Future? A Comprehensive Survey of World Modelscited by 36 here
15 more it builds on, not in this chain
  1. 2018Recurrent World Models Facilitate Policy Evolutioncited by 84 here
  2. 2019When to Trust Your Model: Model-Based Policy Optimizationcited by 30 here
  3. 2020Language Models are Few-Shot Learnerscited by 74 here
  4. 2020ALFWorld: Aligning Text and Embodied Environments for Interactive Learningcited by 45 here
  5. 2023Self-Supervised Learning from Images with a Joint-Embedding Predictive Architecturecited by 32 here
  6. 2023Generative Agents: Interactive Simulacra of Human Behaviorcited by 52 here
  7. 2023Leveraging Pre-trained Large Language Models to Construct and Utilize World Models for Model-based Task Planningcited by 31 here
  8. 2023Reasoning with Language Model is Planning with World Modelcited by 81 here
  9. 2023GAIA-1: A Generative World Model for Autonomous Drivingcited by 30 here
  10. 2023TD-MPC2: Scalable, Robust World Models for Continuous Controlcited by 37 here
  11. 2024WorldCoder, a Model-Based LLM Agent: Building World Models by Writing Code and Interacting with the Environmentcited by 33 here
  12. 2024Genie: Generative Interactive Environmentscited by 66 here
  13. 2024Evaluating the World Model Implicit in a Generative Modelcited by 31 here
  14. 2025Cosmos World Foundation Model Platform for Physical AIcited by 32 here
  15. 2025V-JEPA 2: Self-Supervised Video Models Enable Understanding, Prediction and Planningcited by 30 here

The 24 most cited of 805.

Browse

30 of 486, ranked 6–35

  1. 6WorldCoder, a Model-Based LLM Agent: Building World Models by Writing Code and Interacting with the Environment202433
  2. 7Cosmos World Foundation Model Platform for Physical AI202532
  3. 8Web Agents with World Models: Learning and Leveraging Environment Dynamics in Web Navigation202432
  4. 9Is Your LLM Secretly a World Model of the Internet? Model-Based Planning for Web Agents202432
  5. 10Evaluating the World Model Implicit in a Generative Model202431
  6. 11Leveraging Pre-trained Large Language Models to Construct and Utilize World Models for Model-based Task Planning202331
  7. 12Language Models Meet World Models: Embodied Experiences Enhance Language Models202328
  8. 13Inner Monologue: Embodied Reasoning through Planning with Language Models202224
  9. 14Generating Code World Models with Large Language Models Guided by Monte Carlo Tree Search202423
  10. 15Emergent Linear Representations in World Models of Self-Supervised Sequence Models202322
All 30
  1. 16From Word to World: Can Large Language Models be Implicit Text-based World Models?202520
  2. 17Robust agents learn causal world models202417
  3. 18Emergent World Models and Latent Variable Estimation in Chess-Playing Language Models202417
  4. 19WebEvolver: Enhancing Web Agent Self-Improvement with Coevolving World Model202516
  5. 20From Word Models to World Models: Translating from Natural Language to the Probabilistic Language of Thought202316
  6. 21PoE-World: Compositional World Modeling with Products of Programmatic Experts202514
  7. 22VAGEN: Reinforcing World Model Reasoning for Multi-Turn VLM Agents202514
  8. 23Agent Planning with World Knowledge Model202414
  9. 24Current Agents Fail to Leverage World Model as Tool for Foresight202613
  10. 25S3: Social-network Simulation System with Large Language Model-Empowered Agents202313
  11. 26Code World Models for General Game Playing202512
  12. 27Why Do LLM Agents Fail in Exploring New Environments? A World-Modeling Perspective202512
  13. 28WorldGPT: Empowering LLM as Multimodal World Model202412
  14. 29Making Large Language Models into World Models with Precondition and Effect Knowledge202412
  15. 30MineDojo: Building Open-Ended Embodied Agents with Internet-Scale Knowledge202212
  16. 31CWM: An Open-Weights LLM for Research on Code Generation with World Models202511
  17. 32Language Models, Agent Models, and World Models: The LAW for Machine Reasoning and Planning202311
  18. 33EconAgent: Large Language Model-Empowered Agents for Simulating Macroeconomic Activities202310
  19. 34Neural Theory-of-Mind? On the Limits of Social Intelligence in Large LMs202210
  20. 35Critique of World Model20259