Back to Blog

Gemini Omni 1.2 Flash: 15-Second AI Video With Sound for Marketing

What Gemini Omni 1.2 Flash adds for marketing video: up to 15 seconds with sound, text, photo and clip inputs, 360p to 4K, what changed from 1.1, and credit costs in MITPO.

AINovember 6, 20267 min read
A cellist playing on a rainy tram platform at night as a tram arrives, a frame from a Gemini Omni 1.2 Flash clip

Most marketing video needs two things that older AI video models handled badly: enough length to tell a small story, and sound that belongs to the picture. Gemini Omni 1.2 Flash from Google makes clips of 3 to 15 seconds with synchronized sound, from text, from photos or from a video you already have. Google released it on November 5, 2026, and it replaces Omni 1.1 Flash in MITPO Creative Studio on Starter and above.


Key Takeaways (AEO Direct Summary)

  • Length: 3 to 15 seconds per clip, up from 10 on Omni 1.1, at 360p, 720p, 1080p or 4K and 24 frames per second.
  • Sound: video comes with native audio, so dialogue, ambience and effects are part of the generation.
  • Inputs: text, images and video. Editing accepts a source clip of up to 15 seconds.
  • Price: on Google's developer API, about $0.07 per second at 720p, against about $0.10 on 1.1.
  • In MITPO: 7 credits per second at 720p, so 5 seconds is 35 credits, 10 seconds 70 and 15 seconds 105. Starter and above.

What changed from Omni 1.1

Omni 1.1 FlashOmni 1.2 Flash
Output length3 to 10 seconds3 to 15 seconds
Editing a clipSource up to 10 secondsSource up to 15 seconds
Extending a clipNot listedUp to 30 seconds in total per turn
First and last frameNot listedInterpolation, 3 to 10 seconds
Video price on Google's API$17.50 per 1M tokens$12.00 per 1M tokens
720p cost per secondAbout $0.10About $0.07

The picture quality tiers are the same: 360p, 720p, 1080p and 4K, all at 24 frames per second. Google also plans to send traffic for the 1.1 name to 1.2 on the Gemini API from December 3, 2026, so it is the version to build on.


Length, quality and cost in MITPO

QualityCredits per secondA 5 second clipA 10 second clipA 15 second clip
360p draft3153045
720p73570105
1080p1050100150
4K21105210315

Draft at 360p to check the idea, then render the keeper at 720p or higher. The cost shows before you generate.


How to prompt for sound

Omni makes the sound together with the picture, so describe it the same way you describe the shot. This is our own working method from testing, not vendor guidance:

  1. Name the sounds you want. Rain on a canopy, a soft brake hiss and a low cello line all help. If you leave sound out, the model chooses for you.
  2. Tie each sound to something on screen. A tram arriving is a sound event. So is an engine starting.
  3. One camera idea per clip. A slow orbit or a slow push-in reads better than three moves.
  4. Ask for what stays the same when you start from a photo: "keep the product, the street and the lighting exactly as in the image."
  5. Fix the colour in words. A neutral to cool white balance and film grain keep the look from drifting warm.

In our own tests, 10 second clips at 720p with sound took between 43 and 50 seconds to generate.


Prompt templates

Text
1. Scene with sound (text-to-video, 10 s)
[Subject] [one clear action] in [setting]. Slow cinematic orbit around [subject].
[Something arrives or changes] and its lights sweep across the wet ground.
Natural synced sound: [main sound], [ambient sound], [one effect].
Neutral-to-cool white balance, real film grain, no text.
Text
2. Bring a photo to life (image-to-video, 10 s)
[The motion you want, such as the engine starts and settles into an idle].
Slow cinematic push-in toward [the focal point]. Natural synced sound: [two or three sounds].
Keep [the subject], [the setting] and the lighting exactly as in the image. No text.
Text
3. Change a clip you have (edit, up to 15 s)
Change the setting to [new setting]. Keep [the product], its label and the
camera move unchanged. Keep the original audio feel.

Limits to plan around

  • Extension and first and last frame are not in MITPO yet. Google lists them for 1.2; the studio offers text, photo and clip inputs for now.
  • A single clip tops out at 15 seconds. For longer ads, plan several clips and cut them together.
  • Check faces, hands and small text. They are where AI video most often slips, whatever the model.
  • Sound is generated, not recorded. Review it with headphones before a clip goes out, and replace it if a brand sound matters.

Using Omni 1.2 Flash in MITPO

Pick Gemini Omni Flash 1.2 in Creative Studio, choose 5, 10 or 15 seconds and a quality, then type a prompt, add up to ten reference images, or drop in a clip to change. Plans and limits are on the pricing page, and the Creative Studio docs explain the video options. To compare video models, see AI video models compared, and for the earlier version read the Omni 1.1 guide.


FAQ

What is Gemini Omni 1.2 Flash?

It is Google's video model for generating, editing and extending video with synchronized sound. Google released it on November 5, 2026, as the successor to Omni 1.1 Flash.

How long can an Omni 1.2 Flash video be?

Between 3 and 15 seconds per clip, up from 10 seconds on 1.1. In MITPO a 10 second clip at 720p is 70 credits.

Does Omni 1.2 Flash make sound?

Yes. Video is generated with native audio, and you can describe the sounds you want in the prompt.

What resolutions does it support?

360p, 720p, 1080p and 4K, at 24 frames per second. MITPO offers all four.

How much does Omni 1.2 Flash cost in MITPO?

7 credits per second at 720p on Starter and above, so 5 seconds is 35 credits. Drafts at 360p are 3 credits per second.

Try it

Put this to work in MITPO

Open the demo with no account, or create a free account and start with your own brand.

Share this article