Gemini Omni 1.2 Flash: Google announced Gemini Omni 1.2 Flash on November 5, 2026. The model can generate standard video clips of up to 15 seconds, revise an existing clip with follow-up prompts and produce 720p output at a published API rate of $0.07 per second.
From Prompt to 15-Second Clip: What Changes with Gemini Omni 1.2 Flash
A conceptual DigiOffly illustration of a video workflow: create a draft, refine it in conversation and check the final 15-second output. No interface screenshot or hands-on benchmark is implied.
What Happened
Gemini Omni 1.2 Flash is Google's new video-generation and editing model. Standard clips now extend to fifteen seconds, compared with ten in the previous version. The company cut its published 720p generation rate by 30%. The model is generally available through the Gemini API and Google Cloud, although app features follow their own release schedules. Gemini Omni video generation is billed according to the duration and resolution of the output. The 15-second AI video generation limit applies to standard single-step clips; editing and extension have separate rules.
Who Is Affected & When
Developers can evaluate the Gemini video API in Google AI Studio and the Google Cloud console, provided their project has the necessary permissions and quota. Google also names Flow and Google Ads. In Google Vids, Google says 15-second clips are available at no additional charge with a Google account or Workspace subscription; Google Vids limits still apply and the offer does not include free API credits. Video creation of that length in the Gemini app is rolling out to Google AI Ultra subscribers. Feature access can differ among these interfaces.
What Changes in Practice
Duration limits: standard video versus extensions
The Gemini video length limit for standard generation is fifteen seconds, with whole-second durations starting at three. Video extension uses a different operation that adds three to ten seconds per step. Google Cloud's Enterprise Agent Platform and the Gemini API document different combined-video limits, so developers should check the instructions for their particular endpoint rather than infer a longer single-generation limit. This update does not mean the model can generate an arbitrary-length clip in one request.
Gemini Omni 1.2 Flash pricing: what $0.07 buys
Gemini Omni API pricing starts at a published $0.07 per second of 720p video output, 30% lower than the rate for Omni 1.1 Flash. A 15-second 720p output therefore has approximately $1.05 in video-output charges. A team's AI video generation cost can be higher when extra takes, editing, storage or higher-resolution renders are included. It is not a flat price per finished advertisement.
Refining an existing clip with additional prompts
Gemini Omni multi-turn editing changes an existing clip through follow-up prompts instead of starting each revision from an entirely new scene. Gemini video editing includes requests to adjust lighting, backgrounds or camera movement. The usefulness of video editing by prompt depends on whether revisions preserve identifiable faces, logos and speech. Google has published demonstrations but DigiOffly has not measured the success rate of edits.
Images and audio references: an access caveat
The Gemini Omni video model accepts reference images and short motion clips. In supported Google Cloud workflows, audio cues can guide a scene too, although the company says audio-reference access is limited by allowlisting. These are vendor demonstrations of reference following, not independent evidence that every logo or product label will be accurate.
Planning drafts and upscaled renders
Google describes reviewing a 360p or 720p draft before producing a final 1080p or 4K render. A team can reject an unsuitable scene before paying to create a polished version. A higher-resolution render does not correct an incorrect package label, so inspect the source content before upscaling.
| Consideration | Omni 1.1 Flash | Omni 1.2 Flash |
|---|---|---|
| Maximum standard clip length | Up to 10 seconds | Up to 15 seconds |
| 720p API pricing | Higher prior rate; Google reports a 30% reduction in 1.2 | $0.07 per second of 720p video (Google, November 5) |
| Editing and references | Earlier generation editing tools and reference support | Expanded conversational edits and stronger reference adherence reported by Google |
| Availability | Previous Omni 1.1 Flash model | Generally available via Gemini API and Google Cloud; app rollouts differ |
Why It Matters
A team ordering twenty usable fifteen-second clips would pay roughly $21 for the 720p video-output portion at the advertised rate, before extra attempts or other operations. Gemini Omni pricing should therefore be modeled with an expected rejection rate as well as the number of finished clips. Projects may also be subject to regional availability, quotas, safety policies and Provisioned Throughput options.
Our Take
The longer clip and lower rate make the new version worth evaluating for short advertisements and product demonstrations. A useful pilot would test a product shot, narrated scene and packaging label, then record the number of acceptable results and edits. DigiOffly has not performed those experiments; this is an evaluation proposal based on the product documentation. The linked Google Gemini review covers the broader platform.
Frequently Asked Questions
Is Gemini Omni 1.2 Flash available now?
Google announced general availability on November 5, 2026 for the Gemini API and Google Cloud. Access inside specific apps follows separate rollout and subscription rules.
How long can Gemini Omni videos be?
Standard generation and editing support clips between three and fifteen seconds. Extension adds three to ten seconds per step, with total-length restrictions depending on the Gemini API or Google Cloud endpoint. See Google's linked documentation for the correct extension limit for your integration.
How much does a 15-second Gemini Omni 720p clip cost?
At Google's $0.07-per-second 720p rate, the video-output component for a 15-second clip is approximately $1.05. Other operations and repeated attempts can increase the total bill; this is not a subscription price.
Can Gemini Omni edits preserve logos and product details?
Google reports improved adherence to references, but results should still be reviewed. The launch examples are not an independent guarantee that typography, packaging or brand assets will remain accurate.
Official Sources & Dates
Announcement: 2026-11-05; effective or availability date stated by the vendor: 2026-11-05. Check official updates before purchasing or changing production settings.
- Google official launch announcement, November 5, 2026
- Google Cloud general availability and price announcement, November 5, 2026
- Google Cloud model specifications and operation limits
- Google Cloud video generation documentation
- Google DeepMind model card, November 5, 2026
- Google Cloud documentation: extending videos and per-step limits
- Google Gemini API documentation: Omni video extension limits
Editorial note: DigiOffly verified dated primary sources. This is not a claim of hands-on testing unless independently documented. Prices and access conditions may change.