Google officially announced Gemini Omni Flash at I/O 2026. Now live on Google Flow.

This is the Flash edition of the Omni series. The standout feature is conversational editing:

• Give instructions directly in the chat — "change the background to a Tokyo night scene", "switch to anime style", "extend the action by 3 seconds", "add rain atmosphere and background music"
• AI re-renders the video instantly, with synced audio (dialogue + ambient sound + BGM)

Fast, low barrier, smooth interaction. Purpose-built for short-form video, marketing assets, creative prototyping, and rapid iteration.

Try it (Gemini Advanced / Ultra):

https://labs.google/fx/tools/flow

https://omni-flash.video

Just ran a few prompts — generation speed and consistency both exceeded expectations.

Anyone else already testing it? Would love to compare notes on output quality, character consistency, and prompts that actually work.