Release2026-09-03

Runway released GWM Worlds 2, a research preview that generates playable worlds as they are watched. It produces continuous video and sound that react to a user's inputs. Users define the environment, the characters, the visual style and the physical rules up front. They then steer the world with short typed instructions aimed at a character or at the scene. Camera movement is controlled separately from the character. Sessions have no fixed length, because generation continues from each new input. Runway describes a control format called WorldPrompt, which splits lasting scene context from a timestamped stream of actions. Sessions can also start from an existing eight-second video instead of a still image. Software agents can take the controls, which Runway suggests for testing and training them. Multiple people can each control a different character, or one can direct the world.

What changed

The earlier GWM Worlds simulated environments in real time but without generated audio.

What it unlocks

Playing a generated world live, steering characters and scene events with typed instructions and camera movement.

  • 720p video at 24 fps
  • audio at 48,000 Hz
  • 8s input video can be extended

What you need to act on it

  • contact form submission with business email
  • research preview access

Send this to someone who needs it

Shares the story and its sources. Nothing about you.

What does this mean for your job?

This is the story as everyone gets it. Once a week we send you the version written for your role — what changed, why it matters for the work you actually do, and one thing to try. Free while we tune it.