Runway Releases Praxis‑1 Video Foundation Model, Showcases Leap from Clips to Robots

Runway, best known for its AI‑powered video creation tools, announced on November 11 that its new foundation model "Praxis‑1" is now generally available. While the company has built a strong following among filmmakers and online creators, this launch goes beyond creative workflows: Runway also released research showing how video pretraining with Praxis‑1 can be adapted to robot control.

Praxis‑1 is a large‑scale model trained on vast amounts of video to learn motion, object interactions, and temporal dynamics. It now underpins Runway’s latest product features and partner APIs. In an accompanying research paper, the company demonstrates that policies derived from Praxis‑1 can be fine‑tuned for robotics tasks, allowing robots to learn complex behaviors with less on‑hardware data than traditional methods require.

Runway says it plans to extend Praxis‑1 from creative applications into robotics, simulation, and gaming, positioning the model as a general‑purpose engine for understanding and synthesizing dynamic scenes. A video‑first AI startup pushing into robot control highlights how generative models are quickly becoming a bridge between digital content and the physical world, intensifying competition at that boundary.