mlsanityData to 13 Oct 2026

A Careful Examination of Large Behavior Models for Multitask Dexterous Manipulation

arXiv 2507.05331 ↗First posted 7 Jul 2025cs.ROcited by 127

On the map: Offline RL and imitation learning from data → Robot manipulation imitation learning

Abstract (arXiv)

Robot manipulation has seen tremendous progress in recent years, with imitation learning policies enabling successful performance of dexterous and hard-to-model tasks. Concurrently, scaling data and model size has led to the development of capable language and vision foundation models, motivating large-scale efforts to create general-purpose robot foundation models. While these models have garnered significant enthusiasm and investment, meaningful evaluation of real-world performance remains a challenge, limiting both the pace of development and inhibiting a nuanced understanding of current capabilities. In this paper, we rigorously evaluate multitask robot manipulation policies, referred to as Large Behavior Models (LBMs), by extending the Diffusion Policy paradigm across a corpus of simulated and real-world robot data. We propose and validate an evaluation pipeline to rigorously analyze the capabilities of these models with statistical confidence. We compare against single-task baselines through blind, randomized trials in a controlled setting, using both simulation and real-world experiments. We find that multi-task pretraining makes the policies more successful and robust, and enables teaching complex new tasks more quickly, using a fraction of the data when compared to single-task baselines. Moreover, performance predictably increases as pretraining scale and diversity grows. Project page: https://toyotaresearchinstitute.github.io/lbm1/

Key facts. Cited by 127 papers in the map (data through 13 Oct 2026). Placed in Robot manipulation imitation learning, one of 13 directions in Offline RL and imitation learning from data. First posted on arXiv 7 Jul 2025.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 13 Oct)

Citing papers per year: 2025: 42; 2026: 85.

127 papers.

Sort:

By area:

  1. 2026World Action Models are Zero-shot Policiescited by 438
  2. 2026Causal World Modeling for Robot Controlcited by 305
  3. 2025GR-3 Technical Reportcited by 110
  4. 2025Vision-Language-Action Models for Robotics: A Review Towards Real-World Applicationscited by 82
  5. 2026PointWorld: Scaling 3D World Models for In-The-Wild Robotic Manipulationcited by 79
  6. 2026CaP-X: A Framework for Benchmarking and Improving Coding Agents for Robot Manipulationcited by 76
  7. 2025Large Video Planner Enables Generalizable Robot Controlcited by 67
  8. 2026MEM: Multi-Scale Embodied Memory for Vision Language Action Modelscited by 58
  9. 2025Residual Off-Policy RL for Finetuning Behavior Cloning Policiescited by 43
  10. 2026SceneSmith: Agentic Generation of Simulation-Ready Indoor Scenescited by 42
  11. 2025GR-RL: Going Dexterous and Precise for Long-Horizon Robotic Manipulationcited by 41
  12. 2025RaC: Robot Learning for Long-Horizon Tasks by Scaling Recovery and Correctioncited by 38
38 more
  1. 2025NovaFlow: Zero-Shot Manipulation via Actionable Flow from Generated Videoscited by 38
  2. 2025ResMimic: From General Motion Tracking to Humanoid Whole-body Loco-Manipulation via Residual Learningcited by 36
  3. 2025Much Ado About Noising: Dispelling the Myths of Generative Robotic Controlcited by 36
  4. 2026Native Video-Action Pretraining for Generalizable Robot Controlcited by 32
  5. 2025PolaRiS: Scalable Real-to-Sim Evaluations for Generalist Robot Policiescited by 30
  6. 2026Interactive World Simulator for Robot Policy Training and Evaluationcited by 30
  7. 2025Real-to-Sim Robot Policy Evaluation with Gaussian Splatting Simulation of Soft-Body Interactionscited by 28
  8. 2025In-N-On: Scaling Egocentric Manipulation with in-the-wild and on-task Datacited by 24
  9. 2025The Reality Gap in Robotics: Challenges, Solutions, and Best Practicescited by 22
  10. 2026A Systematic Study of Data Modalities and Strategies for Co-training Large Behavior Models for Robot Manipulationcited by 22
  11. 2026RoboLab: A High-Fidelity Simulation Benchmark for Analysis of Task Generalist Policiescited by 20
  12. 2025Seeing to Act, Prompting to Specify: A Bayesian Factorization of Vision Language Action Policycited by 18
  13. 2026Learning Native Continuation for Action Chunking Flow Policiescited by 18
  14. 2025Do You Need Proprioceptive States in Visuomotor Policies?cited by 16
  15. 2026VLS: Steering Pretrained Robot Policies via Vision-Language Modelscited by 15
  16. 2025DreamControl: Human-Inspired Whole-Body Humanoid Control for Scene Interaction via Guided Diffusioncited by 14
  17. 2025Do You Know Where Your Camera Is? View-Invariant Policy Learning with Camera Conditioningcited by 14
  18. 2025Task adaptation of Vision-Language-Action model: 1st Place Solution for the 2025 BEHAVIOR Challengecited by 14
  19. 2025Reliable and Scalable Robot Policy Evaluation with Imperfect Simulatorscited by 13
  20. 2025TraceGen: World Modeling in 3D Trace Space Enables Learning from Cross-Embodiment Videoscited by 13
  21. 2025REALM: A Real-to-Sim Validated Benchmark for Generalization in Robotic Manipulationcited by 13
  22. 2026VLANeXt: Recipes for Building Strong VLA Modelscited by 12
  23. 2025AnchorDream: Repurposing Video Diffusion for Embodiment-Aware Robot Data Synthesiscited by 11
  24. 2026VisGym: Diverse, Customizable, Scalable Environments for Multimodal Agentscited by 10
  25. 2026RFS: Reinforcement Learning with Residual Flow Steering for Dexterous Manipulationcited by 10
  26. 2025Compose Your Policies! Improving Diffusion-based or Flow-based Robot Policies via Test-time Distribution-level Compositioncited by 9
  27. 2025MoTVLA: A Vision-Language-Action Model with Unified Fast-Slow Reasoningcited by 9
  28. 2026Long-Horizon Manipulation via Trace-Conditioned VLA Planningcited by 9
  29. 2026Hide-and-Seek in Trajectories: Discovering Failure Signals for VLA Runtime Monitoringcited by 8
  30. 2025OSMO: Open-Source Tactile Glove for Human-to-Robot Skill Transfercited by 7
  31. 2026NovaPlan: Zero-Shot Long-Horizon Manipulation via Closed-Loop Video Language Planningcited by 7
  32. 2026Open-H-Embodiment: A Large-Scale Dataset for Enabling Foundation Models in Medical Roboticscited by 7
  33. 2025Human-in-the-loop Online Rejection Sampling for Robotic Manipulationcited by 5
  34. 2026Improving Robotic Generalist Policies via Flow Reversal Steeringcited by 5
  35. 2025A Systematic Study of Large Language Models for Task and Motion Planning With PDDLStreamcited by 4
  36. 2026A Mechanistic Analysis of Sim-and-Real Co-Training in Generative Robot Policiescited by 4
  37. 2026WarmPrior: Straightening Flow-Matching Policies with Temporal Priorscited by 4
  38. 2026HumanoidMimicGen: Data Generation for Loco-Manipulation via Whole-Body Planningcited by 4

50 most cited of 127 All 127 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2026ManiUnit: A Manipulation Skill Dataset and Benchmark for Long-Horizon Tasks ↗
  2. 2026Behavior-Skill: A Fine-Grained Benchmark for Evaluating Vision-Language-Action Policies in Long-Horizon Tasks
  3. 2025Flexible Multitask Learning with Factorized Diffusion Policy
  4. 2024How Generalizable Is My Behavior Cloning Policy? A Statistical Approach to Trustworthy Performance Evaluation
  5. 2026World Models for Robotic Manipulation: A Survey
  6. 2025Dexterous Manipulation through Imitation Learning: A Survey
  7. 2025Learning a Thousand Tasks in a Day
  8. 2026Beyond the Majority: Long-tail Imitation Learning for Robotic Manipulation