Google's latest video model hands the creative decisions back to you: where a scene starts, where it ends, and how long it runs.
Google DeepMind released Gemini Omni 1.1 Flash, a controllable AI video model, on 27 August 2026. The company describes it as production-ready and rolling out, rather than pinning it to a firm general-availability date, so the honest read is that it is released and rolling out from that day. Google AI Plus, Pro and Ultra subscribers can use it immediately in Google Flow, with scene extension also live in the Gemini app.
The headline for anyone who makes video is control. Earlier text-to-video tools gave you one roll of the dice per prompt. This update lets you shape the result. You can extend a scene in increments of up to 10 seconds, to a running maximum of 40 seconds, with steadier visual consistency across the join. You can set the first and last frame so a shot moves the way you intend, in one continuous take. You can feed up to 3 seconds of reference video to hold a character or a style steady across clips.
There is also a cheaper way to work. A draft mode renders at 360p up to 60% faster and at roughly one-third the cost of a 720p render, so you can rough out an idea, get it right, then upscale. Final output is available at 1080p or 4K. On price, Google's own Gemini API pricing page lists video at about $0.10 per second of 720p output under standard rates. The model's identity is set out in the Gemini Omni Flash model card.
What it changes for a firm that makes its own video
For most professional-services firms this is a creative-production tool, not a search or visibility change, and it is fair to treat it that way. If your practice already produces explainer clips, social video or pitch reels, the practical gain is time and cost: the draft-then-upscale workflow means fewer expensive full-resolution renders while you settle on a cut, and keyframe control means less reshooting to get a shot that lands. If you outsource video, it is worth knowing your studio can now iterate faster and cheaper on the same brief.
Beyond your own reach, the model is available through the Gemini API in Google AI Studio and the Gemini Enterprise Agent Platform, so expect it to turn up inside the creative and media-editing tools you already pay for. There is nothing here you need to act on to stay found in search. It is a capability worth filing away for the next time video is on the table.
Frequently asked questions
What is Gemini Omni 1.1 Flash?
It is a controllable AI video-generation model from Google DeepMind, released on 27 August 2026. Its selling point is finer control over the clip: extending scenes, setting start and end frames, feeding short reference clips for consistency, and drafting cheaply before upscaling.
Does Gemini Omni 1.1 Flash require any action from your firm?
No. This is a creative-production tool, not a change to how Google Search or its AI answers rank or cite your site. There is no setting to update and no visibility risk to manage. It is only relevant if you or your studio produce video.
How much does Gemini Omni 1.1 Flash cost?
Google's own pricing page lists video output at roughly $0.10 per second at 720p under standard rates. The draft mode renders at 360p for about one-third of the 720p cost and up to 60% faster, which is aimed at cutting the bill while you iterate. Access to Google Flow requires a Google AI Plus, Pro or Ultra subscription.
How much does Gemini Omni 1.1 Flash cost?
You can extend a scene in increments of up to 10 seconds, to a cumulative maximum of 40 seconds, with improved consistency across each join. Final output is available at 1080p or 4K.
Does Gemini Omni 1.1 Flash help your firm get found in search?
Not directly. The value here is production speed and creative control, not search or AI-answer visibility. If better, faster video helps you win attention on the channels where clients already look, that is a marketing gain rather than a being-found one.
