Gemini Omni 1.2 Flash: The New Guide to Stunning
Google made Gemini Omni 1.2 Flash generally available on November 5, 2026, and the update matters for any team that makes video. The new model generates and edits native video clips up to 15 seconds long, holds characters steady, and costs 30% less than the version it replaces. As a result, short AI video turns from an experiment into a practical weekly workflow for marketers, creators, and small businesses.
This guide explains what changed, what Gemini Omni 1.2 Flash pricing means for a project budget, where the model fits, and where its limits sit. You will also find a four-step workflow, a comparison with Omni 1.1, and answers to common questions.

What Is Gemini Omni 1.2 Flash?
Google DeepMind announced Gemini Omni 1.2 Flash on November 5, 2026, as the new version of its Omni video model, and the company made it generally available the same day. The model turns text, images, and short clips into new video with native audio, and it can edit a clip you already made. In addition, it accepts reference images for characters, products, and styles, plus short motion clips and audio cues, so a product keeps its look throughout.
However, the practical change is length. A Gemini Omni 1.2 Flash 15 second video runs half as long again as the 10-second cap on Omni 1.1, with any length from 3 to 15 seconds in one-second steps. As a result, a product explainer, a social ad, or a short story now fits in a single generation instead of two stitched clips. In addition, Google says the model holds references better and directs smoother camera moves.

Gemini Omni 1.2 Flash Pricing: What a Clip Really Costs
Gemini Omni 1.2 Flash pricing starts at $0.07 per second for 720p video on the paid Gemini API, and Google says that rate is 30% lower than Omni 1.1. Behind the headline, token rates are $1.50 per million input tokens and $12.00 per million output tokens. Therefore a 5-second clip costs about $0.35, a 10-second clip about $0.70, and a 15-second clip about $1.05.
The 30% price cut in plain numbers
The cut matters most at volume. For example, one hundred 15-second clips at 720p cost about $105, while the same batch would have cost roughly $150 at the old implied rate of $0.10 per second. However, the headline rate covers output seconds only. Input tokens, retries, discarded drafts, and upscaling sit outside that number, so a real budget needs room for several attempts per clip.
What the per-second rate leaves out
Drafting at 360p costs less than 720p, so smart teams iterate cheap and upscale only the winner. From December 3, 2026, API calls to gemini-omni-1.1-flash route automatically to 1.2 and bill at 1.2 rates, so there is no reason to build new work on 1.1 now. Meanwhile, Google Vids generates 15-second clips at no additional cost with a Google account or Workspace plan, which makes it the cheapest way to test the model.
| 720p clip | Cost per clip |
|---|---|
| 5-second clip | About $0.35 |
| 10-second clip | About $0.70 |
| 15-second clip | About $1.05 |
Gemini Omni 1.2 Flash vs Omni 1.1: What Changed
The Gemini Omni 1.2 Flash vs Omni 1.1 comparison comes down to four upgrades: longer clips, lower prices, stronger consistency, and conversational editing. In addition, the table below puts the two versions side by side so you can see the gap at a glance.
| Feature | Omni 1.1 Flash | Omni 1.2 Flash |
|---|---|---|
| Max clip length | 10 seconds | 15 seconds |
| 720p price | $0.10 per second | $0.07 per second |
| Editing | Regenerate the clip | Multi-turn chat |
| Consistency | Standard | Stronger |
The upgrade is a swap, not a rebuild. Because the API model ID changes to gemini-omni-1.2-flash and the old ID routes to the new model from December 3, teams can move existing prompts across. On the other hand, results will differ, so run a small comparison batch before you switch a production workflow. In addition, paid-tier rate limits doubled for teams that generate in batches.
Longer Clips, Stronger Consistency
Fifteen seconds sounds small, however it changes what a single clip can hold. A complete product demo, a four-shot explainer, or a short ad now fits in one generation. For example, Google demonstrates this with a 15-second macro explainer inside a beehive, built from one prompt with on-screen callouts that track the motion.
Consistency is the quieter upgrade. The model keeps faces, garments, packaging text, and brand colors steady across camera angles, and it reduces the visual drift that made longer AI clips fall apart. For example, Google shows a virtual try-on sequence built from one model photo and front-and-back shots of three outfits, with the model turning 360 degrees while each garment keeps its cut and drape. Similarly, a first-and-last-frame workflow lets you pin the start and end of a shot while the model fills the motion between them.
Edit in Conversation Instead of Starting Over
Multi-turn editing saves the most time. Instead of re-generating a whole clip when one detail is wrong, you chat with the model and change that detail: swap a background, change the lighting, or switch the aspect ratio. As a result, the rest of the clip stays intact, and the edit applies across the full 15 seconds.
Multi-turn editing for ad variants
Ad teams gain the most here, because one hero clip can become variants for different audiences with a short follow-up prompt each. For example, Google shows one clip re-skinned turn by turn from a striped pavilion tent to an oak forest, a stable, and a stone castle, while the characters and camera framing stay intact. As a result, variant production becomes a conversation, not a rebuild.
Gemini Omni 1.2 Flash for Business: Where It Fits
Gemini Omni 1.2 Flash for business starts with jobs that need many short clips on a deadline: social ads, product explainers, and training snippets. In addition, the model is generally available on the Gemini API and Google Cloud, backed by enterprise infrastructure, Provisioned Throughput, and an imperceptible SynthID watermark. Therefore a team can plan around it as a production tool.
Product videos and virtual try-on
For example, e-commerce teams can turn product photography into catalog spins, virtual try-on clips, and branded video ads. Because the model holds packaging typography and brand colors steady, a product shot stays on-brand. In addition, Artlist, FLORA, invideo, Krea, Picsart, Runway, Photoroom, and QuickFrame already integrate the model, so many teams can use it inside tools they already pay for.
Social ads and explainers
Meanwhile, short-form video drives most social reach, and 15 seconds matches the length of a standard ad spot. A small team can draft a clip in Google Vids at no extra cost, then move to the API as volume grows. In addition, the same workflow serves internal work: onboarding explainers and feature announcements that once needed a shoot day. If your team already writes AI content, the video step now costs about a dollar a clip, so the question changes from whether to try video to how many variants to test.

How to Use Gemini Omni 1.2 Flash: A Simple Workflow
A repeatable workflow keeps costs low and quality high. The four steps below work for a first test and for a weekly production batch.
- Write a prompt that names the subject, the setting, the camera move, and the length.
- Draft at 360p or 720p and review the motion, because a cheap draft exposes a bad prompt early.
- Edit in conversation until the clip is right, since a follow-up prompt costs far less.
- Upscale only the winners to 1080p or 4K.
Draft cheap, upscale the winners
This order is the whole cost discipline. Firstly, cheap drafts catch bad prompts early. Secondly, conversational edits fix details without a rebuild. Thirdly, upscaling runs once per winner, not once per attempt. Finally, the finished clips go straight into Google Vids, Flow, or your ad platform.
Where to Access It: API, Google Vids, Flow, and More
The Gemini Omni 1.2 Flash API is the developer route: prototype in Google AI Studio, then call gemini-omni-1.2-flash from the Gemini API. In addition, paid-tier rate limits doubled at launch, and the model also runs in Flow, Google Ads, and the Gemini app. Gemini Omni 1.2 Flash Google Vids access is the simplest route for non-developers, with 15-second clips at no additional cost.
Availability still varies by product and region. Video editing and avatar creation stay unavailable in the EEA, Switzerland, the United Kingdom, and some US states. Therefore check the access page for your region before you promise a client a delivery date. Google’s flagship language model, Gemini 4 Carbon, arrived a week earlier, and our Gemini 4 Carbon guide explains why it stays limited to cyber defenders. Google’s official announcement lists every access route and the full feature set, and an independent pricing breakdown works through the per-second math.
Honest Limits: What It Still Cannot Do
Gemini Omni 1.2 Flash still struggles with fast sports action, rapid script changes, and complex physics, and Google says work continues there. However, independent benchmarks are also thin: no third-party board had ranked the model in its first days, so the quality claims rest on Google’s own demos. Treat the launch numbers as the maker’s claims, not as verified results.
Fifteen seconds is still a cap, so a finished commercial needs several clips stitched together, and each stitch risks a break. In addition, every output carries a SynthID watermark for provenance. Finally, the per-second price looks tiny, however retries and discarded drafts add up, so track real cost per finished clip, not cost per generation. OpenAI answered November with its own launch, and our GPT-6.1 Terra guide covers the five reasons teams switch fast.
Frequently Asked Questions
How much does Gemini Omni 1.2 Flash cost?
720p video costs $0.07 per second on the paid Gemini API, 30% lower than Omni 1.1. A 10-second clip costs about $0.70 and a 15-second clip about $1.05, before retries.
How long can the videos be?
Clips run from 3 to 15 seconds in one-second steps. First-and-last-frame workflows are limited to 3 to 10 seconds.
Is the model available in Google Vids?
Yes. Google Vids generates 15-second clips at no additional cost with a Google account or Workspace plan.
What is the difference between Gemini Omni 1.2 Flash and Omni 1.1?
Version 1.2 adds 15-second clips, a 30% price cut, stronger consistency, multi-turn editing, and doubled rate limits. From December 3, 2026, calls to 1.1 route to 1.2 automatically.
Can I use my own product photos with the model?
Yes. For example, you can supply character, product, or style images, plus short motion clips and audio cues, and the model holds those references through the clip.
Does the model add a watermark?
Yes. Outputs carry an imperceptible SynthID watermark.
Where can developers access the API?
Prototype in Google AI Studio, then use the model ID gemini-omni-1.2-flash through the Gemini API or Google Cloud.
What can I make with a 15-second video?
A product explainer, a social ad spot, or a virtual try-on sequence. Fifteen seconds fits a complete short message in one generation.
Should small businesses use Gemini Omni 1.2 Flash?
Yes, for short video. Start in Google Vids at no extra cost, then move to the API when you need volume.
Conclusion: Should You Try It?
Gemini Omni 1.2 Flash is the first Google video model built for weekly production rather than demos. As a result, fifteen-second clips, conversational edits, and a $0.07 per-second rate remove the three barriers that kept small teams away from AI video: length, rework, and cost. Start in Google Vids, test one real brief, and measure cost per clip. And if your bottleneck is text agents rather than video, our Mercury 3 guide covers the diffusion model built for agent speed.
[…] integrations are rolling out across OpenRouter, LangChain, Vercel, and LiteLLM. For example, Gemini Omni 1.2 Flash handles video generation, while this model handles the judgments around it. Finally, note that you cannot swap the model ID […]