mlsanityData to 14 Oct 2026

RynnVLA-002: A Unified Vision-Language-Action and World Model

arXiv 2511.17502 ↗First posted 21 Nov 2025cs.ROcited by 74

On the map: World models → Embodied and robot world models

also in 1 more
Abstract (arXiv)

We introduce RynnVLA-002, a unified Vision-Language-Action (VLA) and world model. The world model leverages action and visual inputs to predict future image states, learning the underlying physics of the environment to refine action generation. Conversely, the VLA model produces subsequent actions from image observations, enhancing visual understanding and supporting the world model's image generation. The unified framework of RynnVLA-002 enables joint learning of environmental dynamics and action planning. Our experiments show that RynnVLA-002 surpasses individual VLA and world models, demonstrating their mutual enhancement. We evaluate RynnVLA-002 in both simulation and real-world robot tasks. RynnVLA-002 achieves 97.4% success rate on the LIBERO simulation benchmark without pretraining, while in real-world LeRobot experiments, its integrated world model boosts the overall success rate by 50%.

Key facts. Cited by 74 papers in the map (data through 14 Oct 2026). Placed in Embodied and robot world models, one of 11 directions in World models. Builds on 52 of its 80 references that are in the map. First posted on arXiv 21 Nov 2025.

Builds on

52 of 80 references.

Closest ancestors · 8

Same direction, newest first

  1. 2025World-Env: Leveraging World Model as a Virtual Environment for VLA Post-Training29 Sep
  2. 2025IRL-VLA: Training an Vision-Language-Action Policy via Reward World Model7 Aug
  3. 2025WorldVLA: Towards Autoregressive Action World Model26 Jun
  4. 2025FLARE: Robot Learning with Implicit World Modeling21 May
  5. 2024GenRL: Multimodal-foundation world models for generalization in embodied agents26 Jun
  6. 2024iVideoGPT: Interactive VideoGPTs are Scalable World Models24 May
  7. 20243D-VLA: A 3D Vision-Language-Action Generative World Model14 Mar
  8. 2023Diffusion Policy: Visuomotor Policy Learning via Action Diffusion7 Mar

Other references · 44

  1. 2022Scalable Diffusion Models with Transformerscited by 6,054
  2. 2022Flow Matching for Generative Modelingcited by 5,702
  3. 2023RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Controlcited by 3,080
  4. 2024OpenVLA: An Open-Source Vision-Language-Action Modelcited by 2,887
  5. 2024π0: A Vision-Language-Action Flow Model for General Robot Controlcited by 2,407
  6. 2025π0.5: a Vision-Language-Action Model with Open-World Generalizationcited by 1,761
  7. 2019Dream to Control: Learning Behaviors by Latent Imaginationcited by 1,653
  8. 2018World Modelscited by 1,513
  9. 2025GR00T N1: An Open Foundation Model for Generalist Humanoid Robotscited by 1,225
  10. 2020Mastering Atari with Discrete World Modelscited by 1,007
  11. 2025Fine-Tuning Vision-Language-Action Models: Optimizing Speed and Successcited by 883
  12. 2025Cosmos World Foundation Model Platform for Physical AIcited by 864
32 more
  1. 2024Genie: Generative Interactive Environmentscited by 747
  2. 2025FAST: Efficient Action Tokenization for Vision-Language-Action Modelscited by 665
  3. 2025SmolVLA: A Vision-Language-Action Model for Affordable and Efficient Roboticscited by 530
  4. 2025CoT-VLA: Visual Chain-of-Thought Reasoning for Vision-Language-Action Modelscited by 451
  5. 2024CogACT: A Foundational Vision-Language-Action Model for Synergizing Cognition and Action in Robotic Manipulationcited by 422
  6. 2025UniVLA: Learning to Act Anywhere with Task-centric Latent Actionscited by 410
  7. 2023An Embodied Generalist Agent in 3D Worldcited by 365
  8. 2024TinyVLA: Towards Fast, Data-Efficient Vision-Language-Action Models for Robotic Manipulationcited by 339
  9. 2022Transformers are Sample-Efficient World Modelscited by 328
  10. 2024Latent Action Pretraining from Videoscited by 321
  11. 2024TraceVLA: Visual Trace Prompting Enhances Spatial-Temporal Awareness for Generalist Robotic Policiescited by 304
  12. 2025X-VLA: Soft-Prompted Transformer as Scalable Cross-Embodiment Vision-Language-Action Modelcited by 303
  13. 2025MolmoAct: Action Reasoning Models that can Reason in Spacecited by 197
  14. 2025DexVLA: Vision-Language Model with Plug-In Diffusion Expert for General Robot Controlcited by 195
  15. 2025ThinkAct: Vision-Language-Action Reasoning via Reinforced Visual Latent Planningcited by 156
  16. 2025VLA-Adapter: An Effective Paradigm for Tiny-Scale Vision-Language-Action Modelcited by 145
  17. 2022TransDreamer: Reinforcement Learning with Transformer World Modelscited by 143
  18. 2023Transformer-based World Models Are Happy With 100k Interactionscited by 139
  19. 2025NORA: A Small Open-Sourced Generalist Vision Language Action Model for Embodied Taskscited by 127
  20. 2023Language Models Meet World Models: Embodied Experiences Enhance Language Modelscited by 124
  21. 2025ChatVLA: Unified Multimodal Understanding and Robot Control with Vision-Language-Action Modelcited by 113
  22. 2024WorldDreamer: Towards General World Models for Video Generation via Predicting Masked Tokenscited by 89
  23. 2024HiRT: Enhancing Robotic Control with Hierarchical Robot Transformerscited by 75
  24. 2025VQ-VLA: Improving Vision-Language-Action Models via Scaling Vector-Quantized Action Tokenizerscited by 70
  25. 2024LLaRA: Supercharging Robot Learning Data for Vision-Language Policycited by 52
  26. 2025VLAS: Vision-Language-Action Model With Speech Instructions For Customized Robot Manipulationcited by 51
  27. 2024Diffusion Transformer Policycited by 41
  28. 2023Paragraph-to-Image Generation with Information-Enriched Diffusion Modelcited by 39
  29. 2025EO-1: An Open Unified Embodied Foundation Model for General Robot Controlcited by 31
  30. 2022DreamingV2: Reinforcement Learning with Discrete World Models without Reconstructioncited by 29
  31. 2025GEVRM: Goal-Expressive Video Generation Model For Robust Visual Manipulationcited by 27
  32. 2025RynnVLA-001: Using Human Demonstrations to Improve Robot Manipulationcited by 25

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 74.

74 papers.

Sort:

By area:

  1. 2026Fast-WAM: Do World Action Models Need Test-time Future Imagination?cited by 349
  2. 2026Cosmos 3: Omnimodal World Models for Physical AIcited by 165
  3. 2026World Action Models: The Next Frontier in Embodied AIcited by 50
  4. 2026World Model for Robot Learning: A Comprehensive Surveycited by 46
  5. 2026World Guidance: World Modeling in Condition Space for Action Generationcited by 36
  6. 2026When Vision Overrides Language: Evaluating and Mitigating Counterfactual Failures in VLAscited by 30
  7. 2026OA-WAM: Object-Addressable World Action Model for Robust Robot Manipulationcited by 28
  8. 2026JEPA-VLA: Video Predictive Embedding is Needed for VLA Modelscited by 24
  9. 2026Dream-Tac: A Unified Tactile World Action Model for Contact-Rich Robot Manipulationcited by 19
  10. 2026WALL-WM: Carving World Action Modeling at the Event Jointscited by 17
  11. 2026MaskWAM: Unifying Mask Prompting and Prediction for World-Action Modelscited by 16
  12. 2026World-Value-Action Model: Implicit Planning for Vision-Language-Action Systemscited by 14
38 more
  1. 2026VLANeXt: Recipes for Building Strong VLA Modelscited by 12
  2. 2026Kinema4D: Kinematic 4D World Modeling for Spatiotemporal Embodied Simulationcited by 12
  3. 2026World Action Models: A Surveycited by 12
  4. 2026MMaDA-VLA: Large Diffusion Vision-Language-Action Model with Unified Multi-Modal Instruction and Generationcited by 10
  5. 2026ProgressVLA: Progress-Guided Diffusion Policy for Vision-Language Robotic Manipulationcited by 10
  6. 2026HarmoWAM: Harmonizing Generalizable and Precise Manipulation via Adaptive World Action Modelscited by 10
  7. 2026LIBERO-Safety: A Comprehensive Benchmark for Physical and Semantic Safety in Vision-Language-Action Modelscited by 10
  8. 2026Hide-and-Seek in Trajectories: Discovering Failure Signals for VLA Runtime Monitoringcited by 9
  9. 2026DeMaVLA: A Vision-Language-Action Foundation Model for Generalizable Deformable Manipulationcited by 9
  10. 2026Kairos: A Regret-Aware Native World-Action Model Stack for Physical AIcited by 9
  11. 2026DSWAM: A Dual-System World Action Foundation Model for Fine-Grained Robot Manipulationcited by 8
  12. 2026NoiseGate: Learning Per-Latent Timestep Schedules as Information Gating in World Action Modelscited by 7
  13. 2026One Token Per Frame: Reconsidering Visual Bandwidth in World Models for VLA Policycited by 6
  14. 2026The Great March 100: 100 Detail-oriented Tasks for Evaluating Embodied AI Agentscited by 5
  15. 2026PACT-WAM: Predicting Actions and Visual Foresight with Compact Temporal Encoding for Robot Manipulationcited by 5
  16. 2026Brain-WM: Brain Glioblastoma World Modelcited by 5
  17. 2026Pre-VLA: Preemptive Runtime Verification for Reliable Vision-Language-Action and World-Model Rolloutscited by 5
  18. 2026ACE-Brain-0.5: A Unified Embodied Foundational Model for Physical Agentic AIcited by 5
  19. 2026Recursive Belief Vision Language Action Modelscited by 4
  20. 2026OxyGen: Unified KV Cache Management for VLA Inference under Multi-Task Parallelismcited by 4
  21. 2026Sword: Style-Robust World Models as Simulators via Dynamic Latent Bootstrapping for VLA Policy Post-Trainingcited by 4
  22. 2026Escaping the Diversity Trap in Robotic Manipulation via Anchor-Centric Adaptationcited by 4
  23. 2026Qantara: Bridge-Flow Training for Multi-Paradigm JEPA Controlcited by 4
  24. 2026SG-WAM: Self-Guided World Modeling in Geometry-Aware Policy Spacecited by 4
  25. 2026Foresight Without Seeing: Latent Futures for World Action Modelscited by 4
  26. 2026PROSPECT: Unified Streaming Vision-Language Navigation via Semantic--Spatial Fusion and Latent Predictive Representationcited by 3
  27. 2026RoVLA: Multi-Consistency Constraints for Robust Vision-Language-Action Modelscited by 3
  28. 2026AEGIS: A Backup Reflex for Physical AIcited by 3
  29. 2026ActProbe: Action-Space Probe for Early Failure Detection of Generative Robot Policiescited by 3
  30. 2026World Pilot: Steering Vision-Language-Action Models with World-Action Priorscited by 3
  31. 2026Attacking the Trusted Imagination: Oracle-Level Integrity Attacks on Imagine-then-Act World Modelscited by 3
  32. 2026Keep the Future, Drop the Rollout: RIFT for World Action Modelscited by 3
  33. 2026Audio-Visual Intelligence in Large Foundation Modelscited by 2
  34. 2026From Imagined Futures to Executable Actions: Mixture of Latent Actions for Robot Manipulationcited by 2
  35. 2026VLA-Pro: Cross-Task Procedural Memory Transfer for Vision-Language-Action Modelscited by 2
  36. 2026PointAction: 3D Points as Universal Action Representations for Robot Controlcited by 2
  37. 2026VISTA: Vision-Grounded and Physics-Validated Adaptation of UMI data for VLA Trainingcited by 2
  38. 2026DVG-WM: Disentangled Video Generation Enables Efficient Embodied World Model for Robotic Manipulationcited by 2

50 most cited of 74 All 74 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2026Do World Action Models Generalize Better than VLAs? A Robustness Study
  2. 2026Think Like a World Model, Act Like a VLA: Distilling World-Model Representations into Compact Robot Policies
  3. 2026Towards VLA-Dreamer: Refining VLA Behavior Using World Models
  4. 2026Towards Practical World Model-based Reinforcement Learning for Vision-Language-Action Models
  5. 2026GigaBrain-0.5M*: a VLA That Learns From World Model-Based Reinforcement Learning
  6. 2025An Anatomy of Vision-Language-Action Models: From Modules to Milestones and Challenges
  7. 2026WAVE-0: Bridging VLA Policies and WAMs with Lightweight Latent Future Prediction
  8. 2025LoHoVLA: A Unified Vision-Language-Action Model for Long-Horizon Embodied Tasks