Reward AI Releases OM-1: A Robot Policy Trained on Human Demonstrations Only, With No Teleoperation or On-Robot Data

Reward AI has released OM-1, a general-purpose robotic manipulation policy trained exclusively on human demonstrations captured via wearable sensors. By eliminating the need for complex teleoperation or on-robot training data, the model allows industrial arms and humanoid robots to learn new tasks from fewer than three minutes of demonstration. This approach aims to simplify the training process for complex physical interactions by leveraging human-like movement patterns directly.