Gemini Omni Can Now Direct AI Video and Finish in 4K

Google is shipping Gemini Omni 1.1 Flash, a production update for its video model that can extend a scene, lock first and last frames, and upscale the result to 4K, Google said Thursday.

When Google dropped Omni at I/O in May, the pitch was conversational video from any mix of text, images, audio, and clips. Today’s update is the control layer. The model now reads up to 10 seconds of prior footage when you say continue, up from the last second, and you can stack those extensions in 10-second chunks to a 40-second cap.

You can also hand it a first and last frame and let it fill the motion between them, which is the path for orbits, zooms, and loops. Drafts can render at 360p, which Google says is up to 60 percent faster and a third of the cost of its standard 720p. Final output can go to 1080p or 4K. Video references in a prompt can run up to three seconds.

Adobe has it in Firefly. Figma Weave, Runway, and GMI Cloud are already using Omni Flash in production. Gemini API docs.

That sits next to the Google Photos tool that uses Omni to fix badly lit videos. Omni 1.1 is live today for Google AI Plus, Pro, and Ultra subscribers in Flow, with scene extension in the Gemini app.