GWM-1 Avatars / Characters
GWM-1 Avatars is the model direction behind Runway Characters. The Characters service creates a conversational video persona from a single image, with configurable voice, personality, knowledge and actions. Runway launched the service in March 2026.
Separate a live character from a rendered clip
The Characters launch describes live conversation with an expressive visual persona. The Dev catalog also lists character-video generation from a text script or audio file. Those are different product workflows, even though both relate to the avatar model family.Runway Characters launchRunway Dev model catalog
The appearance does not define the character’s knowledge
Runway lets developers configure the character’s knowledge and conversational actions separately from its appearance. A realistic face or expressive delivery does not establish that an answer is correct. Supply the information and permitted actions that the application actually needs.Runway Characters launch
Practical guidance. For a product guide, review the information source and action permissions before adding the avatar. Then inspect interruption behavior, audio clarity and the visual response during both speaking and listening.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | API, Provider product |
| Model identifiers | gwm1_avatars |
| API model identifier | gwm1_avatars for generated character videos.Runway Dev model catalog |
| Character setup | A single reference image; the launch announcement says no fine-tuning is required.Runway Characters launch |
Related models
A performance-transfer model that applies speech, expression and movement to a character image or video.
Launch date not confirmedAPI · Provider product