Gemini Omni 1.1 Flash (Generally Available)

Google's video generation and editing model, out of preview and into general availability as gemini-omni-1.1-flash.

Generates video from text with audio, from a single image, or from multiple reference images of the same subject
Extends an existing clip by 3 to 10 seconds, and fills the gap between a first and last frame to build a transition
Edits are made in plain language through the Interactions API, which tracks the chain of edits so each one builds on the last
Outputs at 360p, 720p by default, 1080p, or 4K, in vertical or landscape
Editing or extending video you uploaded yourself is unavailable in the European Economic Area, Switzerland, and the UK, though editing model-generated video works everywhere
The older preview endpoint shuts off September 30, 2026