Reward AI trains OM-1 with human demos.
A wearable records motion, touch, and images.
Egocentric data remains key for robot learning.
Now it’s being paired with UMI-style systems.
The same model controls arms and humanoids 🦾
Introducing OM-1, our first robot foundation model, zero-shot generalizing to any robot: table-top arms, industrial arms and humanoids.
- learned directly from human manipulation data
- no teleop/robot data
- close to human-level dexterity and efficiency
- multi-robot collab