Google’s Gemini Omni Update Makes 15-Second AI Video Available to Developers

Google has lengthened the clips its Gemini Omni video model can make, moving from a 10-second ceiling in Omni 1.1 Flash to as much as 15 seconds in Omni 1.2 Flash. Announced on November 5, the update also lets a user refine a generated clip through follow-up instructions rather than beginning again. For a creator trying to adjust a background or change a camera angle, that is a practical shift: the first result can become a starting point instead of a disposable draft.

The company says Omni 1.2 Flash is generally available through the Gemini API and on Google Cloud for developers and businesses. Google lists 720p generation at $0.07 per second, which it describes as 30 percent below the prior Flash model’s rate. A full 15-second clip at that rate would cost $1.05 for video generation before any other usage or processing charges. Clips can be specified in one-second increments from three to 15 seconds. Lower-resolution drafting and later upscaling are also supported, according to Google.

Consumer access is not identical to developer access. Google says 15-second generation is available in Google Vids at no additional charge with a Google account or Workspace subscription, while the rollout in the Gemini app is for Google AI Ultra subscribers. Flow and Google Ads are among the other products receiving the model. Users should check the actual tool and account before assuming a feature in the API is already switched on in their own app or region.

Google also claims stronger adherence to reference images, short motion clips and other prompts, and more natural-sounding dialogue. Those are the company's performance descriptions, not independently verified guarantees. Its Cloud announcement notes that audio references require allowlisting. It says outputs include an imperceptible SynthID watermark. The result may be especially useful to small creative teams making product demonstrations, short explainers or several versions of an advertisement, though realistic-looking generated video still needs clear disclosure when viewers might mistake it for documentary footage.

This update is about production flexibility, not an unlimited free video service. The model’s longer ceiling and editing workflow are available now in some channels, while consumer access and costs vary by product and subscription. That distinction matters more than a polished demo when deciding whether the release will change a reader’s own work.