RoboCasa365: A Large-Scale Simulation Framework for Training and Benchmarking Generalist Robots
2026-03-04 • Robotics
RoboticsArtificial IntelligenceMachine Learning
AI summaryⓘ
The authors created RoboCasa365, a big virtual test with 365 common household tasks in 2,500 kitchen-like settings, to better measure how well robots can learn many jobs at once. They included lots of example actions from both real humans and computers to help train robots. By running experiments with current robot learning methods, the authors explored how different factors like task variety and data size affect robots' ability to do many tasks. Their work helps researchers understand what is important for making robots that can work in many everyday situations.
robot learningbenchmarkhousehold tasksmobile manipulationmulti-task learningrobot foundation modelslifelong learninggeneralizationsimulation environment
Authors
Soroush Nasiriany, Sepehr Nasiriany, Abhiram Maddukuri, Yuke Zhu
Abstract
Recent advances in robot learning have accelerated progress toward generalist robots that can perform everyday tasks in human environments. Yet it remains difficult to gauge how close we are to this vision. The field lacks a reproducible, large-scale benchmark for systematic evaluation. To fill this gap, we present RoboCasa365, a comprehensive simulation benchmark for household mobile manipulation. Built on the RoboCasa platform, RoboCasa365 introduces 365 everyday tasks across 2,500 diverse kitchen environments, with over 600 hours of human demonstration data and over 1600 hours of synthetically generated demonstration data -- making it one of the most diverse and large-scale resources for studying generalist policies. RoboCasa365 is designed to support systematic evaluations for different problem settings, including multi-task learning, robot foundation model training, and lifelong learning. We conduct extensive experiments on this benchmark with state-of-the-art methods and analyze the impacts of task diversity, dataset scale, and environment variation on generalization. Our results provide new insights into what factors most strongly affect the performance of generalist robots and inform strategies for future progress in the field.