1X

The world model, trained on thousands of hours of video data, can predict how environments and objects respond to robot actions.

What stands out most in these demonstrations is not just what NEO does, but how it does it—silently, efficiently, and with a level of dexterity that appear convincingly human-like.