Gemini Omni 1.2 Flash: 15-Second AI Video, 30% Cheaper

Updated: November 20265 min readBy the CreatorSprout team

Launch coverage based on Google’s November 5, 2026 announcement and its published API pricing, as reported by multiple outlets, we have not tested Gemini Omni 1.2 Flash hands-on. Capability and quality claims below are Google’s, not independent test results. No affiliate links in this article.

Omni 1.2 Flash is the first version of Google’s video model to reach general availability, backed by Google Cloud’s enterprise infrastructure, and it arrives as a direct answer to the two complaints creators had about Omni 1.1: clips were too short, and fixing one thing in a clip meant re-rolling the whole thing. The 15-second ceiling and multi-turn editing attack both. For creators, this is the update that makes Google’s video model usable for a real Short or Reel segment rather than a teaser fragment, but the pricing story splits in two depending on whether you pay per second through the API or per credit in Flow, and that split is the detail most launch coverage glosses over.

What was announced

Google announced Gemini Omni 1.2 Flash on November 5, 2026, in a launch post by Anish Nangia, a product manager at Google DeepMind, and made it generally available the same day through the Gemini API and Gemini for business on Google Cloud. The release has three parts: longer native clips (15 seconds, up from 10), multi-turn conversational editing, and stronger reference fidelity, faces and products stay more consistent from the reference image through the generated clip. The price cut is the fourth part: 720p video drops to $0.07 per output second, 30% below Omni 1.1.

Omni 1.1, released in August 2026, capped a single clip at 10 seconds, so anything longer had to be built by extending a clip segment by segment. Omni 1.2 Flash is also the first Omni model with a general-availability label, which matters for production use: Google is offering it with Provisioned Throughput (reserved capacity) for enterprise customers on Google Cloud. Outputs carry an imperceptible SynthID watermark, consistent with Google’s approach to labelling AI-generated content.

15-second clips: what actually changed

3 to 15 seconds, in one-second steps. You pick the exact length instead of accepting a fixed 10-second cap. A 15-second clip covers a full Shorts hook, a product demo beat, or a B-roll segment without stitching.
First-and-last-frame workflows stay shorter. Workflows that pin both a starting and an ending frame are limited to 3 to 10 seconds, the 15-second ceiling applies to standard generation and editing.
Consistency is the quieter upgrade. Google says reference images are reproduced more faithfully, so a face or a product holds its look through the clip. For creators, this is the difference between a usable brand shot and one that drifts into a stranger by the third second, but it is a vendor claim until independent testing exists.
Audio is part of the package. Clips come with synchronised audio, and the model supports audio references, though audio-reference input is gated behind allowlisting rather than generally open, so most creators will not have it on day one.

Multi-turn editing: fix the clip, don’t re-roll it

The feature that changes the workflow most is not the extra five seconds, it is conversational editing. Instead of regenerating a whole clip when one detail is wrong, you send a follow-up message and the model revises the same scene: change the lighting, swap an object, tighten a movement, keep everything else. Google describes this as editing designed to revise the same scene rather than restart from a blank prompt.

For creators this is the practical difference between a tool and a slot machine. Re-rolling a 10-second clip until the background is right burns credits and patience; a follow-up edit that keeps the good parts and changes one thing is how real editing works. The honest limit is the same as with every multi-turn video tool: each edit round is still a generation, and continuity across many rounds is exactly what needs hands-on testing before you trust it on a client deliverable.

Pricing and the Flow credit catch

API pricing, $0.07 per second for 720p. A full 15-second 720p clip costs about $1.05. Google’s published token basis is $12 per million tokens for standard video output, with 720p calculated at 5,792 tokens per second. The API is available on the paid Gemini API tier, the free tier does not include this model.
1080p and 4K have no published per-second rate. Google has published a per-second price for 720p only. If you need higher resolutions, check the pricing page before budgeting, do not assume the 30% cut applies evenly.
Flow credits went the other way. In Google Flow, a 10-second 720p clip now costs 30 credits, up from 15 credits on Omni 1.1. And 15-second clips, Extend and upscaling in Flow need a Google AI Plus, Pro or Ultra plan, free accounts stay at 10 seconds. So the same model is cheaper per second through the API and more expensive per clip in Flow’s credit system. Which route is cheaper for you depends entirely on where you generate.
Google Vids is the free-entry route. Google says 15-second clips in Vids cost nothing extra with a Google account or Workspace subscription, with a full rollout over one to three days from November 5.
Retirement dates to plan around. Omni 1.1, Veo 3.1 Fast and Veo 3.1 Quality retire in Flow on December 5, 2026. API requests to gemini-omni-1.1-flash will automatically route to 1.2 from December 3, 2026, and be billed at the newer model’s rates. Google lists a retirement date of November 5, 2027 or later for Omni 1.2 Flash itself, so treat the model ID as temporary and keep a migration plan in mind.

Where you can use it

Developers: Gemini API and Google AI Studio (model ID gemini-omni-1.2-flash), and Vertex AI on Google Cloud, including Provisioned Throughput for high-volume production work.
Google Flow: text, frames and reference-to-video generation, video editing and multi-turn edits. 15-second clips, Extend and upscaling need a paid Google AI plan; free accounts stay at 10 seconds.
Google Vids: 15-second generation rolling out across listed Workspace and consumer editions, at no additional cost with a Google account or Workspace subscription.
Gemini app: 15-second video is rolling out to Google AI Ultra subscribers.
Google Ads: available for ad creative workflows.

What it means for creators

A 15-second clip is a usable unit. Ten seconds was a teaser; fifteen seconds is a complete Shorts hook, a product beat, or a B-roll segment. If you build faceless or B-roll-heavy content, this is the first Omni version that fits a real editing timeline without stitching.
Multi-turn editing saves the most expensive resource: re-rolls. Try it on a clip you already like and change one element with a follow-up prompt. If the scene holds, your iteration cost drops sharply, that is the feature to test first, before the extra seconds.
Price the route, not the model. API at $0.07 per second is genuinely cheaper, but Flow credits doubled per clip. If you generate in Flow on a paid plan, your per-clip cost went up, not down. Run your own numbers on the route you actually use before assuming the 30% cut applies to you.
Vids is the no-cost way to try it. If you have a Google account or Workspace subscription, 15-second clips in Vids cost nothing extra, that is the lowest-risk place to test whether the quality and consistency claims hold for your content.
Don’t rebuild on day one. Omni 1.1 retires in Flow on December 5, so you will move to 1.2 whether you plan to or not, but keep your existing video workflow until you have seen a multi-turn edit survive your own quality bar.

The honest caveats

We haven’t tested it. Everything here comes from Google’s announcement and its published API pricing as reported by multiple outlets. The consistency improvements, the editing quality and the real-world cost per usable clip are Google’s claims, not results we measured.
The price cut is not universal. The 30% cut is an API per-second figure for 720p. Flow’s credit cost per clip doubled, and no per-second rate is published for 1080p or 4K. Read the price as “cheaper through the API at 720p,” not “cheaper everywhere.”
15 seconds is still short. A longer clip is not a finished commercial. Continuity, audio quality, rights and factual accuracy still need human review, and longer-form video still needs stitching, Extend, or a different tool.
Audio references are gated. The allowlisted audio-reference input means most creators cannot use it yet. Do not plan a workflow around a feature you may not have access to.
Rollouts are staggered. The Gemini app rollout is limited to Ultra subscribers for now, and Vids visibility takes one to three days. If 15-second generation is missing from your account, that is the staged rollout, not a bug.

Verdict

Gemini Omni 1.2 Flash is a genuinely useful update, and the editing, not the extra seconds, is the reason to care. Fifteen-second clips make the model fit a real creator timeline, and the 30% API price cut is real at 720p. But the honest version of the excitement is the usual one: multi-turn editing is the feature that could change your workflow, the Flow credit hike is the catch that could change your bill, and neither is proven until you run your own clip through it. Try it in Vids first, where it costs nothing extra, test one multi-turn edit on a real scene, and keep your existing video tool until the results survive your own quality bar.

FAQ

What is new in Gemini Omni 1.2 Flash?

Gemini Omni 1.2 Flash, released November 5, 2026, generates and edits native video up to 15 seconds (up from 10 seconds on Omni 1.1), adds multi-turn conversational editing, improves how faithfully reference images are reproduced, and cuts 720p pricing to $0.07 per second, 30% below Omni 1.1.

How long can Gemini Omni 1.2 Flash videos be?

You can generate and edit clips from 3 to 15 seconds, choosing the length in one-second steps. First-and-last-frame workflows are limited to 3 to 10 seconds. In Google Flow, 15-second clips need a Google AI Plus, Pro or Ultra plan; free accounts stay at 10 seconds.

How much does Gemini Omni 1.2 Flash cost?

Through the Gemini API, 720p video costs $0.07 per output second, so a full 15-second clip is about $1.05. That is 30% less than Omni 1.1. Google has not published a per-second rate for 1080p or 4K. In Google Flow, a 10-second 720p clip costs 30 credits, up from 15 credits on Omni 1.1, and in Google Vids 15-second clips cost nothing extra with a Google account or Workspace subscription.

Where can I use Gemini Omni 1.2 Flash?

It is generally available through the Gemini API and Google AI Studio, Vertex AI on Google Cloud, Google Flow, Google Vids, Google Ads, and the Gemini app, where 15-second video is rolling out to Google AI Ultra subscribers. Audio-reference input is allowlisted rather than generally open.

Is Gemini Omni 1.1 going away?

Yes. Google says Veo 3.1 Fast, Veo 3.1 Quality and Omni Flash 1.1 retire in Flow on December 5, 2026, and API requests to gemini-omni-1.1-flash will automatically route to 1.2 from December 3, 2026. Google lists a retirement date of November 5, 2027 or later for Omni 1.2 Flash itself.