Gemini Omni 1.2 Flash: 15-Second Video and a Creator’s Guide

Checked November 9, 2026. On November 5, Google announced Gemini Omni 1.2 Flash, a video generation model with clips up to 15 seconds and conversational editing. Here is what the announcement means for creators, where to look for access, and how to test it responsibly. This is a guide to Google’s announcement, not an independent MSL test.

日本語版はこちら

What changed?

Duration: Google says native clips can run from 3 to 15 seconds in one-second increments, compared with up to 10 seconds in Omni 1.1.
Conversational editing: ask for follow-up changes to elements such as the background, lighting or aspect ratio.
References: Google describes combining character, product or style images with short motion clips and audio cues, and claims better consistency across scenes.

These are Google’s descriptions, not guarantees. Compare product details and on-screen text with the source, and review the full clip for changes in faces, hands or objects.

Where can you use it?

Google announced general availability in the Gemini API and Gemini for business, with rollout across products including Flow and Google Vids. In the Gemini app, 15-second video is rolling out to Google AI Ultra subscribers. Availability, settings and usage limits can vary by region, account and organization; check your own interface before planning a project.

For an individual: check the video-generation screen, access and remaining allowance in the app you use.
For developers: check the current model, input support, quotas and pricing in Gemini API and Google AI Studio.
For teams: establish who may upload materials, who pays, and which organization rules apply.

What might it cost?

Google’s launch post lists 720p API generation at $0.07 per second. At that rate, a 15-second clip is $1.05 in generation charges; four such attempts would be $4.20. These are simple estimates under the stated conditions, excluding taxes, conversion, additional processing and further attempts. They are not subscription prices or prices for every resolution. Check the current official pricing and the actual interface, including editing and upscaling terms, before spending.

Try a fictional 15-second event promo

This practice prompt is an illustrative example written by MSL, not an actual generated result. Start with fictional materials and adapt the settings to the product available to you.

Create a 15-second vertical video introducing a fictional photography exhibition called “Shapes of Light.” Move slowly from the entrance into a bright gallery, showing abstract photographs on the walls and the backs of adult visitors viewing them. Use smooth camera motion and natural light. Do not show real artworks, logos or on-screen text. End on a clear wall with room for a title to be added later.

Decide where the video will appear; verify aspect ratio, duration, allowance and cost.
Use reference images only when you have the necessary rights, and identify the colors and composition that must stay consistent.
After watching the first draft, request one change while stating what to preserve: “Brighten the background; keep the people’s positions and camera movement.”
Add dates and captions later in an editor using approved copy.
Play the result on a phone and check the opening, movement, cuts, sound, accuracy and rights.

Five checks before using the result

Consistency: do faces, hands, product shapes, colors and patterns change across frames?
Readability: can viewers see the subject on a small screen, and is there room for captions?
Sound: check voices, lip sync, unwanted sounds and rights to audio.
Rights and privacy: confirm permission for reference images, likenesses, audio and publication; avoid uploading confidential material without approval.
Total effort: record attempts, editing time and cost. Where product accuracy is essential, consider real footage or conventional editing.

Google notes that fast sports action and complex physics remain areas for improvement. Start with a small test and keep a human final review.