On November 5, 2026, Google released Gemini Omni 1.2 Flash, making it generally available on the Gemini API and Google Cloud. It’s the company’s newest AI video model, and it makes clips up to 15 seconds long, with 720p video output priced at $0.07 per second. Google says that price is 30% lower than Omni 1.1, according to its official announcement.
That $0.07 figure is the video-output rate only. Google’s Gemini API pricing page bills input separately ($1.50 per million tokens for text, image, or video input) and text output at $9.00 per million tokens. Google puts a complete 15-second 720p clip at about $1.04 in video-output cost alone, before those other charges.
If you’ve ever tried to cram a product pitch into a 10-second AI clip, you know the pain. You end up stitching three clips together and hoping the coffee mug doesn’t change color between them. The short answer: Omni 1.2 Flash gives you more runway, lower-cost output options, and edits without regenerating the entire clip. Whether it beats other AI video tools is a separate question, and nobody has answered it independently yet.
What Happened With Gemini Omni 1.2 Flash
Google’s announcement post lists these changes:
- Longer clips: You pick a length from 3 to 15 seconds, in one-second steps. Omni 1.1 topped out at 10 seconds.
- Multi-turn editing: Ask for a follow-up change, like swapping a product, changing the lighting or weather, or switching the aspect ratio. You don’t have to regenerate the whole clip.
- Better reference adherence: Feed it photos of a person, product, or style, plus short motion clips (up to three seconds) and audio cues as references. Google says faces, garments, packaging typography, and brand colors stay consistent across camera angles.
- Less drift: Google says character and product consistency is improved from first frame to last, including when you extend a clip or fill in between start and end keyframes, though it doesn’t detail exactly how morphing is measured.
- Draft cheap, then upscale: Generate drafts at 360p or 720p, fix them with edits, and only pay to upscale the winner to 1080p or 4K.
Think of it like printing a rough proof on cheap paper before sending the final to the photo lab. You can keep early drafts cheaper, then pay the higher output rate only when you upscale the version you want to keep.
Who Can Use It
The model is generally available through the Gemini API (paid tier) and Google Cloud, which is the developer route, and in Gemini for business. Google says it’s also rolling out to Google Vids, where 15-second clips come at no additional cost with a Google account or Google Workspace subscription, plus Flow and Google Ads.
In the consumer Gemini app, 15-second video creation is currently rolling out to Google AI Ultra subscribers. The announcement doesn’t give a timeline for other Gemini plans, and because it’s a staged rollout, it may not appear on every eligible account right away.
One catch on price: Google bills API video output by tokens (5,792 tokens per second of 720p video at $12 per million output tokens), and the $0.07 figure is the effective per-second equivalent, not a flat per-clip price. Input tokens, text output, and every edit or retry that generates new video add further charges on top of that rate.
Google’s pricing page lists the effective video-output rates at other resolutions too: about $0.023 per second at 360p, $0.104 per second for 1080p with upsampling, and $0.209 per second for 4K with upsampling. The developer model ID is gemini-omni-1.2-flash, and Google’s pricing page says API requests to gemini-omni-1.1-flash will automatically route to 1.2 Flash, billed at 1.2 rates, starting December 3, 2026. There’s no free API tier for the model.
How It Stacks Up
Google only compares Omni 1.2 Flash against its own predecessor:
| Feature | Gemini Omni 1.1 | Gemini Omni 1.2 Flash |
|---|---|---|
| Max clip length | 10 seconds | 15 seconds (3 to 15, one-second steps) |
| Video-output price (720p, Gemini API) | About $0.10 per second | About $0.07 per second (output rate; excludes input/text charges) |
| Edit without regenerating | Not listed by Google | Yes, multi-turn editing |
| Draft resolutions | Not listed by Google | 360p or 720p, upscale to 1080p or 4K |
Sources: Google blog, November 5, 2026; Gemini API pricing page
Against OpenAI’s Sora or other AI video tools, there’s no fair comparison to make. We couldn’t find any independent benchmarks, reviewer tests, or head-to-head cost comparisons against OpenAI’s Sora or Google’s own Veo line. Every quality and consistency claim above comes from Google alone.
Google also admits there are weak spots. It says development is ongoing to improve things like fast sports action, rapid script changes, and complex physics.
The Reaction
Public reaction is still forming, so there’s no meaningful pattern to report yet.
Our Take
Verdict: worth a look for marketers and small shops making simple product videos. Hold off if your ideas involve lots of fast motion.
The people who’ll get the most out of this are ad creators, social media marketers, and anyone with a product catalog. Fifteen seconds is enough for what Google calls a full social ad spot in one generation, though longer ads with narration or disclosures may still need more than one clip. Reference adherence targets exactly what sellers need, like a 360-degree product spin or a virtual try-on that keeps the logo readable.
I’d skip it for now if your project involves fast sports action or complex physics. Google itself has flagged those as areas it’s still improving.
What to do next: if you have Google AI Ultra, watch for 15-second videos in the Gemini app as the rollout reaches you. Or look for it in Google Vids if your work account has it. Start with a 720p draft before paying for anything bigger, and budget for input and text charges on top of the per-second output rate.
Wrapping Up
Gemini Omni 1.2 Flash gives you 50% more clip length, edits that don’t start from scratch, and a cheaper per-second output price on paper. That’s a real upgrade for anyone stitching short promo clips together. But “longer and cheaper” only tells you what Google measured, and nobody outside Google has tested whether it looks better than the competition. I’d treat it as a promising draft tool until independent reviews show up.