Google’s Gemini Omni 1.2 Flash makes 15-second AI videos, three free a month

Google‘s newest video model, Gemini Omni 1.2 Flash, generates clips up to 15 seconds long with sound and dialogue, and anyone with a Google account can now make three of them a month at no cost in Google Vids, the company’s video editor. The model is also cheaper for developers and lets you fix a clip by asking for changes in conversation instead of generating it again. For anyone who tried AI video and gave up after a few seconds of footage, the extra length is the gap between a fragment and a finished scene.

In practice, that means a product spot, a short explainer or a social clip can come out of a single generation, instead of three short pieces that never quite match when stitched together. You pick the length in one-second steps, from three to fifteen seconds. You can also feed the model photos of a character, a product or a visual style, motion clips of up to three seconds and audio cues, and Google says faces, clothes, packaging text and brand colors now hold steady from the first frame to the last.

The bigger change for everyday use is editing by conversation. Instead of rewriting a prompt and hoping, you ask for one change at a time: a new background, a different product, other lighting, rain instead of sun, a vertical crop. In Google’s own demonstration, one clip of two armored characters moves from a striped pavilion to an oak forest, a stable, a market fair and a stone castle while their performances and camera framing stay the same. Google also claims clearer lip-sync and better handling of scenes where several characters talk.

For developers and businesses, the price drops 30% compared with Omni 1.1 Flash, to $0.07 per second of 720p video through the Gemini API. A full 15-second clip therefore costs about $1.05, roughly what a 10-second clip cost on the previous version. Drafts can be made at 360p or 720p and only the keepers upscaled to 1080p or 4K. It is also the first time Gemini Omni is generally available on Google Cloud, which is what lets apps such as Picsart, Photoroom, Runway and Krea build it into their own products.

The free part is narrower than it sounds. Three clips a month adds up to 45 seconds of footage, enough to experiment but not to run a channel. In the Gemini app, the 15-second videos are rolling out only to subscribers of Google AI Ultra, the company’s most expensive plan, which is not sold in every country, and Google did not publish a country list for the free Vids tier. Google itself admits the model still struggles with fast sports action, rapid script changes and complex physics. Audio references, the feature that lets you supply a voice sample, sit behind an allowlist on Google Cloud.

There is also the question of what viewers will see. Every Omni clip carries SynthID, Google’s watermark, but it is imperceptible: detection tools can read it, while someone scrolling a feed sees nothing that marks the video as generated. As faces stay consistent and lip-sync improves, that difference matters more than it did when AI video still looked like AI video.

Google announced Gemini Omni 1.2 Flash on November 5, 2026, and says it is rolling out now across the Gemini API, Google Cloud, Vids, Flow and Google Ads, with the Gemini app following for Ultra subscribers. Fifteen seconds is still shorter than most social videos, but it is exactly the length of an ad, and that is the market Google priced this model for.

Tags: Google, Gemini, Generative AI, AI video, Gemini Omni 1.2 Flash, Google Flow