From the source
Runway has released Praxis-1, an open weight world action model that turns its video pretraining into control for real robots, built to work around the scarcity of real-world robot data.
- The model converts a 27B-parameter video foundation model into a world model that predicts the future in short chunks, with a small action expert reading the video model's internal state as it imagines what comes next.
- Praxis-1 fine-tuned on DROID is the top world action model and the top open-source robot foundation model on the RoboLab-120 leaderboard, according to Runway, and reaches a competitive success rate against the previous state-of-the-art world action model on a real Franka arm.
- The backbone has 26.9B parameters, of which about 2.8B are active for any token, and the action expert is a dense 16-layer transformer of about 314M parameters.
- On each call the model takes an instruction, the latest camera frames and the robot's current pose, and returns a predicted chunk of future video along with 1.6 seconds of actions: 24 action rows at 15 Hz.
- Runway says Praxis-1 adapts to various embodiments, from single arms to bimanual robots and mobile manipulators, including embodiments it has not seen in training data.






