The short version: Google opened its Gemini Omni Flash video model to developers on June 30, and the number that matters for a small business owner is the price tag: about $0.10 per second of finished video, so a 10-second product clip or social ad runs roughly a dollar. The bigger shift is not the price. It is that you make the video by talking to it, the same way you would ask an assistant to redo a line in an email, instead of learning editing software.
What Gemini Omni Flash actually does
Gemini Omni Flash is a video generation and editing model that Google DeepMind released in public preview through the Gemini API, Google AI Studio, and its Enterprise Agent Platform. Feed it a text prompt, a product photo, or a short reference clip, and it produces a 10-second, 720p video. What sets it apart from the AI video tools that showed up over the past two years is the editing loop: you can ask for a change in plain English, such as “make the background darker” or “have the bottle rotate slower,” and it revises the same clip while keeping the subject and motion consistent, rather than generating a brand new video from scratch each time.
Google is pairing it with a second new model, Nano Banana 2 Lite, which turns a text prompt into a still image in about four seconds for roughly three cents. The intended workflow is to generate a product image first, then hand it to Omni Flash to animate. A five-second animated product shot, start to finish, lands around fifty cents in raw generation cost.
How much does an AI video for small business actually cost?
Google’s published rate is $0.10 per second of 720p output on the paid API tier, which comes out to $17.50 per million video output tokens. There is no free tier for this model, so the meter starts on the first clip. A 10-second video, currently the maximum length the model supports, costs about a dollar. That is the same per-second rate Google charges for Veo 3.1 Fast, its other video model, but Omni Flash adds the conversational editing that Veo does not have.
Compare that to what a short product or promo video has historically cost a small business: a freelance videographer for a single afternoon shoot, a stock footage license, or a subscription to a template-based video app that still requires someone on staff to learn the interface. None of those options run a dollar a clip, and none of them let you fix a mistake by typing a sentence.
What can a small business actually make with it?
Three uses fit squarely inside what a 10-second, 720p clip can do well:
- Product demos. Turn a photo of a menu item, a piece of furniture, or a retail product into a short clip that shows it from another angle or in motion, useful for a website or a marketplace listing.
- Social ads. Instagram Reels, TikTok, and YouTube Shorts already favor short vertical or square video over static images. A one-dollar clip is cheap enough to test five variations of an ad before spending on paid placement.
- Email and landing page teasers. A short looping clip embedded in a promotional email tends to hold attention longer than a photo, without the production overhead of a full commercial shoot.
The considered take: the price is the headline, but the real unlock is who can now do this work. Video production used to require either a specialist or a real time investment learning software like Premiere or CapCut. A conversational interface means the owner who already writes their own social captions can direct a video the same way, describing what they want changed instead of clicking through timeline tools. That is a meaningfully lower floor, not just a lower price.
What are the catches?
A few real limits are worth knowing before you build a workflow around this. Clips top out at 10 seconds and 720p, so this is not a replacement for a full commercial or a 4K hero video on a homepage. Google caps effective iteration at a few sequential edits before quality on complex scenes starts to drift, so treat it as three or four rounds of revision, not unlimited tinkering. Content filters block generating real people’s names or likenesses, so this will not produce a video of your actual staff or a specific customer. And every output carries Google’s SynthID watermark, an invisible marker that survives compression, which is a feature, not a flaw: it means your customers and platforms can verify the clip was AI generated if that ever matters.
The bottom line
None of this replaces a real product photoshoot or a well-produced brand video when the budget and the moment call for one. What it does is remove the excuse for not having any video at all, which is the more common problem for small businesses competing on feeds that reward motion over static posts. If you have never made a video because you did not have the time, the software skill, or the budget, this is the first tool priced and built simply enough to change that math. Related developments on the same theme worth reading: how Gemini now reads your Google Business Profile directly, what Gemini’s real-time call translation means for customer-facing teams, and how Google restructured its AI subscription pricing for SMBs earlier this year.
Frequently Asked Questions
What is Gemini Omni Flash?
It is a video generation and editing model from Google DeepMind, released in public preview on June 30, 2026. It creates 10-second, 720p videos from text, images, or short reference clips, and lets you revise the result through natural-language conversation instead of editing software.
How much does it cost to make a video with Gemini Omni Flash?
Google charges $0.10 per second of finished video on the paid API tier. A 10-second clip, the current maximum, costs about $1.00. There is no free tier, so cost starts with the first generation.
Can I use Gemini Omni Flash directly, or do I need a developer?
As of this preview, Omni Flash is accessed through the Gemini API, Google AI Studio, or the Enterprise Agent Platform, which means either a developer on your team or a third-party app built on top of the API. Google has not announced a standalone consumer product yet, so most small businesses will reach it through tools built by other vendors in the coming months.
What are the biggest limitations right now?
Clips are capped at 10 seconds and 720p resolution, quality holds up for only a few rounds of sequential edits before it can drift on complex scenes, and the model will not generate real people’s names or likenesses. Every clip also carries an invisible SynthID watermark marking it as AI generated.
If a dollar buys you a usable product video, what would you actually make first: a demo of your best-selling item, a five-second social teaser, or something else entirely? Tell us in the comments.
