Quick Answer
Gemini Omni 1.2 Flash, generally available from 5 November 2026, generates and edits video clips up to 15 seconds long at $0.07 a second for 720p video on the paid Gemini API, which works out to about $1.04 for a full 15-second clip. That price makes a small video clip service viable: a $59 package of three 15-second clips costs about $3.30 in video output tokens, so the margin is in your time, your prompting skill and your revisions, not in the API bill. The costs below are illustrations on Google's published rates, not a promise of income.
Key takeaways:
- Gemini Omni 1.2 Flash launched on 5 November 2026 and is generally available on the Gemini API and Gemini for business.
- Clips now run from 3 to 15 seconds, in one-second steps, up from a 10-second maximum, with multi-turn conversational editing and stronger consistency.
- Google cut the price 30 percent from Omni 1.1, to $0.07 a second for 720p video, or about $1.04 for a full 15-second clip.
- The API is paid tier only, with no free tier, and video output is billed at $12 per million tokens, at 5,792 tokens per second of 720p video.
- In Google Flow, 15-second clips, Extend and upscaling need a Plus, Pro or Ultra plan, and Flow charges credits per second, which is a different price from the API.
- Every output carries a SynthID watermark, and from 3 December 2026, requests to Omni 1.1 automatically route to 1.2 rates.
- The package maths in this article uses $29, $59 and $129 packages as labelled illustrations, with API costs of about $1.10, $3.30 and $9.50, and no income is promised.
What changed with Gemini Omni 1.2 Flash?
Gemini Omni 1.2 Flash is Google's video generation and editing model, and the 5 November 2026 release is a practical upgrade rather than a new idea. The headline changes are three: longer clips, better editing, and a lower price. Clips now run up to 15 seconds, in one-second steps from 3 seconds, where Omni 1.1 stopped at 10 seconds. Multi-turn conversational editing lets you change one detail with a follow-up instruction instead of re-rendering the whole clip. And Google says the model follows references more strongly and holds consistency better across edits, which is the difference between a usable product video and a clip where the product changes shape halfway through.
The price cut is the lever for a service business. Google cut the price 30 percent from Omni 1.1, to $0.07 a second for 720p video. That is the rounded figure Google quotes. The exact token maths behind it is $12 per million video-output tokens, with 720p video calculated at 5,792 tokens per second, so a 15-second clip uses 86,880 video-output tokens and costs $1.04256, or about $1.04. A 10-second clip costs about $0.70, and a 5-second clip costs about $0.35. Input is billed separately at $1.50 per million tokens for text, image or video, and text output, including thinking tokens, is billed at $9 per million tokens. A typical clip's full bill lands near $1.05 to $1.25 once inputs and thinking tokens are included, as an illustration, because prompt sizes and thinking vary.
The table below shows what clips cost at 720p on the API, using Google's published token rates. These are video-output costs only, before the small input and thinking charges.
| Clip length | Video-output tokens | Video-output cost | | 3 seconds | 17,376 | About $0.21 | | 5 seconds | 28,960 | About $0.35 | | 10 seconds | 57,920 | About $0.70 | | 15 seconds | 86,880 | About $1.04 |
Two dates matter for anyone building on this model. First, it is generally available now on the Gemini API and Gemini for business, with Provisioned Throughput available for heavier use. Second, from 3 December 2026, API requests to gemini-omni-1.1-flash will automatically route to gemini-omni-1.2-flash and be billed at the newer model's rates, so there is no need to migrate code, and no reason to build a service on the old model now.
What exactly would you sell?
The service is simple: short AI video clips for small businesses that need video but cannot afford a shoot. The customers are the same ones who already buy social media graphics and short ads: local shops, Etsy sellers, coaches, restaurants, real estate agents, and online sellers launching a product. They do not need a 60-second brand film. They need a 15-second clip of their product in use, a looping background for a landing page, or three variations of a short ad to test on Instagram and TikTok.
Sell clips, not minutes. A 15-second product clip, a 10-second logo animation, or a 5-second social loop are all real deliverables a client can post the same day. The workflow is: the client sends product photos or a short brief, you generate a first clip, you edit it with follow-up instructions until it looks right, you upscale the keeper if the package includes it, and you deliver the file with a note about the watermark. Multi-turn editing is what makes this workflow work, because the first generation is rarely the final one, and the ability to say "keep everything, but make the background warmer" without starting over saves both time and tokens.
Three package shapes cover most small clients. The Starter is one 15-second clip for a single product or announcement. The Standard is three 15-second clips, usually three angles or three messages for the same product, so the client can test which one performs best. The Premium is eight 15-second clips plus two rounds of edits across the set, for a client running a small campaign. Those packages are illustrations to show the maths, not a price list to copy blindly. Your local market, your portfolio and your turnaround time should set your real prices.
What does it cost you per package?
The API cost is the smallest part of the bill, which is exactly why this service works. The table below uses the published $0.07 a second rate for 720p video output, plus a small allowance for input and thinking tokens, and treats the package prices as illustrations.
| Package | Price (illustration) | Clips | Video-output cost | Full API cost (illustration) | Gross margin before your time | | Starter | $29 | 1 x 15 seconds | About $1.04 | About $1.10 | About $27.90 | | Standard | $59 | 3 x 15 seconds | About $3.12 | About $3.30 | About $55.70 | | Premium | $129 | 8 x 15 seconds | About $8.32 | About $9.50 | About $119.50 |
The full API cost column adds a small allowance for input tokens at $1.50 per million and text output at $9 per million, which covers prompts, reference images and thinking tokens on a typical job. It also assumes one generation per clip. In practice you will generate two or three versions of some clips before the client approves one, so a realistic allowance is to double the video-output cost for revisions. Even doubled, a Standard package costs about $6.60 in API charges against a $59 illustration price, and a Premium package costs about $19 against a $129 illustration price. The API is not the constraint. Your time is.
Time is the honest cost to quote to yourself. A Starter package, including the client call, the brief, two or three generations, one edit round and delivery, takes about 1 to 2 hours when you are new, and less once you have templates. A Standard package takes about 2 to 3 hours. A Premium package takes about 4 to 6 hours across two days, because client feedback waits on the client. Those are illustrations, and your speed will differ, but they set the right expectation: at two Standard packages a month, the illustration revenue is $118, the illustration API cost is about $6.60, and the illustration time is 4 to 6 hours. Nobody gets rich on two packages a month. The service only becomes real income with repeat clients and a steady flow, and any article that promises otherwise is lying to you.
Which route should you use: API, Flow, or Vids?
There are three ways to make the clips, and they have different prices, so picking the wrong one quietly eats your margin.
The API route is the cheapest per clip and the most flexible, and it is the route the package maths above assumes. You pay per token, you control the prompts, and you can automate generation for repeat clients. It is paid tier only, with no free tier, so you need a billed Google Cloud project. Use this route when you are selling clips as a service and you want the lowest cost per clip.
Google Flow is the visual workspace, and its pricing works differently. Flow charges credits per second, not tokens. For Omni 1.2 Flash, generating or editing up to 10 seconds costs 3 credits a second at 720p, and seconds 11 to 15 cost 6 credits per extra second for Plus, Pro and Ultra users. That means a 10-second 720p clip costs 30 credits, and a 15-second clip costs 60 credits. A 10-second clip cost 15 credits on Omni 1.1, so the same clip costs twice as many Flow credits on 1.2, even though the API price fell. Free accounts stay at 10 seconds, and 15-second clips, Extend and upscaling need a Plus, Pro or Ultra plan. Google AI Pro costs $19.99 a month after its free period, as previously verified. Do not mix up Flow credits and API tokens when you quote a client a price, because they are two different price lists for the same model.
Google Vids is the third route, and it is the cheapest for a particular job: Google says 15-second clips cost nothing extra in Vids with a Google account or Workspace subscription. That makes Vids the right tool for internal videos, presentations and simple clips where the Vids editor is enough, but it is not a general-purpose client service tool, and it does not replace the API for automated or high-volume work.
The practical rule is simple. Use the API for client work you bill by the clip. Use Flow when you want a visual workspace and you have a Plus, Pro or Ultra plan, and check the credit cost in Flow's prompt-box settings before you generate, because Google says Flow limits and credit costs can change at any time. Use Vids for your own content and for simple client videos where the Vids editor fits the job.
What are the limits and traps?
Five limits decide whether this service works for a given client. First, every output carries a SynthID watermark, Google's invisible marker that identifies the video as AI generated. Tell clients about it before they pay, not after delivery, because a client who discovers it afterwards will feel misled, and a client who knows in advance will usually not care.
Second, 15 seconds is the maximum. A client who wants a 30-second story, a testimonial, or a product walkthrough needs either two clips edited together or a different tool. Sell 15 seconds honestly, and do not promise longer.
Third, the API has no free tier for this model. You cannot test the service for free on the API, so budget a few dollars of test generations before your first client, and do not promise a free sample that costs you API charges you did not plan for.
Fourth, Flow credit costs rose for the same clip even as the API price fell, from 15 credits to 30 credits for a 10-second 720p clip. If you build your service in Flow instead of on the API, your costs follow the credit price list, not the $0.07 a second figure. Confirm the live credit cost before you quote.
Fifth, consistency is better but not perfect. Google says Omni 1.2 Flash follows references more strongly, but a product, a face or a logo can still drift across edits, especially in longer clips. Check every delivered clip frame by frame before you send it, and keep the client's original product photo as the reference so you can spot drift early.
How do you find the first clients?
Start where the product is already visual. Message three local shops, Etsy sellers or online sellers whose products photograph well, and offer one free 15-second clip in exchange for a testimonial and permission to show it in your portfolio. One free clip costs you about $1.10 in API charges and an hour of time, and it gives you a real example to show the next ten clients. Do not offer ten free clips. One per niche, three niches maximum, then charge.
The pitch is one sentence: "I make 15-second product videos from your photos, $29 for one clip, delivered in two days." That sentence works because it names the deliverable, the price and the deadline, and it lets the client say yes or no without a meeting. Follow up once, politely, and move on if there is no reply. A client who needs video will answer. A client who does not need it will not, and no amount of follow-up changes that.
Keep the first month small and honest. Aim for two or three paying clients, deliver on time, collect a testimonial from each, and put the clips in a simple portfolio page. If those clients come back for a second order, you have a service. If they do not, ask what was missing, fix it, and try the next three clients. That is how a small service grows: one good clip, one happy client, one repeat order at a time.
Sources: Google blog, "Gemini Omni 1.2 Flash: 15-second video and lower pricing", 5 November 2026; Google Gemini API pricing page, checked 6 November 2026; Google Cloud announcement, "Gemini Omni 1.2 Flash now generally available", 5 November 2026; Unite.AI, "Google's Gemini Omni 1.2 Flash Reaches General Availability", 5 November 2026; 7minAI, "How to Make 15-Second Videos with Gemini Omni 1.2 Flash in Flow, Vids and the Gemini App", checked 6 November 2026; ExpertBeacon, "Gemini Omni 1.2 Flash: Cost, Access And Editing Limits", checked 6 November 2026; Google AI Pro price of $19.99 a month as previously verified.
Related reading: TopUseAI Blog, Sell AI Video Service on AIVideoFree, Pika H3 Flex
Frequently Asked Questions
What is Gemini Omni 1.2 Flash?
It is Google's video generation and editing model, generally available from 5 November 2026, making clips from 3 to 15 seconds long with multi-turn conversational editing.
How much does a 15-second clip cost on the API?
About $1.04 for the video-output portion at 720p, based on $12 per million video-output tokens at 5,792 tokens per second. Input and thinking tokens add a few cents, so a typical clip costs near $1.05 to $1.25 in total.
Is there a free tier for Gemini Omni 1.2 Flash on the API?
No. The model is available on the paid Gemini API tier only, with no free tier. In Flow, free accounts stay at 10-second clips, and 15-second clips need a Plus, Pro or Ultra plan.
How much should I charge for a 15-second AI video clip?
That depends on your market, but the illustrations in this article use $29 for one clip, $59 for three clips, and $129 for eight clips plus two edit rounds. The API cost is about $1.10, $3.30 and $9.50 for those packages, so your time and revisions are the real cost.
Do Gemini Omni 1.2 Flash videos have a watermark?
Yes. Every output carries a SynthID watermark, and you should tell clients about it before they pay.
What is the difference between the API price and Flow credits?
The API charges $0.07 a second for 720p video, while Flow charges credits per second: 3 credits a second up to 10 seconds at 720p, and 6 credits per extra second for seconds 11 to 15. A 10-second 720p clip costs 30 Flow credits on Omni 1.2, double the 15 credits it cost on Omni 1.1.



