Google announced Gemini Omni 1.2 Flash on November 5, 2026. The video model now supports clips up to 15 seconds long, more consistent use of reference images and conversational edits that can refine an existing clip. For creators making AI Shorts, product videos or animated characters, the update is particularly relevant because longer shots and repeatable visual details are two recurring production challenges.
What changed in Gemini Omni 1.2 Flash?
According to Google’s announcement, the new version increases native video duration from 10 seconds in Omni 1.1 to 15 seconds, with selectable durations from three to fifteen seconds in one-second steps. Google also says the model follows character, product and style references more closely, including across changes in camera angle.
The company highlights multi-turn conversational editing: users can request changes to a clip, such as adjusting the background, lighting or framing, without rewriting the entire original brief. It also describes improvements to dialogue, lip-sync, camera moves and first-to-last-frame continuity. These are Google’s reported improvements, not independently verified results from AI Revolution.
Why this matters for AI video creators
A 15-second generation can cover a complete short scene rather than forcing an immediate cut after a few seconds. For an animated character, a useful test would be to keep the same reference image, voice and outfit while requesting a short action and a camera move. A second test could use follow-up editing to change only the background, then check whether the face, clothing and motion remain stable.
For product content, a practical workflow would be to supply a product reference, generate a rotating or moving shot, then ask for a different setting. The important question is whether the product’s shape, labels and colors survive each edit. We have not run these tests yet; they are suggestions for evaluating the update rather than claims about its real-world reliability.
Availability, pricing and limitations
Google says Gemini Omni 1.2 Flash is generally available through the Gemini API and Google Cloud. The announcement also lists access through Google Flow and Google Vids, while availability in the Gemini app is rolling out to Google AI Ultra subscribers. Access and limits can therefore differ by product.
Google’s announced Gemini API price for 720p generation is $0.07 per second, which it describes as 30% lower than Omni 1.1. That figure is specific to the cited API offering and resolution; it should not be assumed to apply to every interface, output resolution or third-party provider.
The company acknowledges that fast sports action, rapid script changes and complex physics still need improvement. Creators should also expect to review individual frames rather than assuming reference consistency is flawless.
What to test first
If the model is available in your workflow, start with a short, controlled comparison: use the same character reference and prompt in your previous model and Omni 1.2 Flash, record duration and generation cost, then compare face consistency, motion, audio and unwanted scene changes. Save the exact prompts and outputs so the results can be reproduced.
For related practical advice, read our guide to keeping the same character across AI videos and our camera movement prompt tests.
Official source: Google — Gemini Omni 1.2 Flash announcement (November 5, 2026).
Editorial note: This news report summarizes Google’s official announcement. The cover is an original editorial illustration, not an actual model output. AI Revolution has not independently tested Gemini Omni 1.2 Flash for this report.