AW-0

We jointly model code, future video, and robot actions as three coupled generative streams within a unified mixture-of-transformers architecture.

The model is conditioned on the current multi-view observations, the language instruction, the robot state, and a compact memory of the program slices generated so far.

Real-world hybrid evaluation is limited to YAM, as code-as-policy systems for dexterous hands are lacking, and the system does not yet learn online from its own interactions.