Google DeepMind

Generative AI · Video Generation · Developer Tools

Launch

Google DeepMind launches Gemini Omni 1.1 Flash with 40-second scenes, 4K video

The advertised length is stitched from shorter takes and the crisp resolution is upscaled, not native, details that matter for billing and client promises.

• Google DeepMind released Gemini Omni 1.1 Flash on August 27, 2026, a production update to its video generation model available via the Gemini API in Google AI Studio and the Gemini Enterprise Agent Platform.
• Omni is Google's native multimodal video model, combining text, image, audio and video input with stateful conversational editing so creators can request changes without re-uploading footage.
• Scene extension now reads up to 10 seconds of prior footage instead of just the final frame, letting a clip grow in 10-second increments to a cumulative 40 seconds.
• New first-and-last-frame control lets developers pin a shot's opening and closing images so the model generates the motion between them, and up to three 3-second reference clips can steer style and character consistency.
• A new 360p draft tier costs about $0.03 per second, roughly a third of 720p pricing, while a full 40-second 4K clip runs about $12; both 1080p and 4K outputs are upscaled from a 720p render rather than generated natively.
• Google is retiring the gemini-omni-flash-preview endpoint that launched June 30 on September 30, 2026, giving developers a five-week window to migrate to the GA model.
• The release reframes generative video as a directable production tool rather than a prompt lottery, addressing the consistency and cost problems that kept earlier models out of professional workflows.