# Gemini Omni 1.2 Flash: what it is and what changes for SME video > Gemini Omni 1.2 Flash generates and edits native video. What Google claims, what is missing, and how to assess it without getting carried away. Source: Giallo Studio (giallostudio.it) · Page: https://giallostudio.it/en/blog/gemini-omni-flash-video-pmi · Published: 2026-11-05 · Updated: 2026-11-05 Gemini Omni 1.2 Flash is a Google model announced on the company's official blog: according to the material, it brings native video generation and editing with 15-second clips, greater consistency and a price that is 30% lower. This is stated on [Google's presentation page](https://blog.google/innovation-and-ai/technology/developers-tools/introducing-gemini-omni-1-2-flash/), which is the only source this article relies on. In simple terms, a model of this kind takes a written description (and, depending on the case, some starting material) and produces or changes a short film. For an entrepreneur or a professional, the useful question is not whether it is impressive, but whether it can replace or lighten work that today costs time and money: a product video, a social media story, an animated presentation. Here we separate what Google states from what the material does not say. The most important point is that the starting price is missing, and without it the claimed saving cannot be turned into a quote. Before spending, then, you need official rates and a test on your own case.

In summary

## What did Google announce with Gemini Omni 1.2 Flash? The announcement is brief and should be read for what it is. Google's blog, on the page titled "Gemini Omni 1.2 Flash: Fast and cost-effective with longer clips", summarizes three novelties: native video generation and editing with 15-second clips, stronger consistency and a cost that is 30% lower. All three claims are reported by the [official source](https://blog.google/innovation-and-ai/technology/developers-tools/introducing-gemini-omni-1-2-flash/). The title itself speaks of longer clips. The material available to us does not say longer than what, so we will not make comparisons with earlier versions that we cannot document. If you want to understand how to think when a provider releases a new version of a model family, you can read the [comparison between two Gemini Flash versions](https://giallostudio.it/en/blog/gemini-3-6-flash-vs-3-5-flash), which shows how to approach the move from one model to another. To frame the broader context, it is also useful to see the guide on [what Gemini is and how it works](https://giallostudio.it/en/blog/what-is-gemini-ai): it explains the family of models this novelty belongs to, with no technical knowledge required. ## What does it mean that video generation and editing are native? The source uses the adjective "native" for generation and editing. The available text does not explain it, so what follows is our own cautious reading, not a statement from Google: it means that the two operations are functions of the model itself, and not a chain of separate tools lined up one after another. Why does it matter? Because generating a video and editing it are different jobs. In the first case you start from scratch; in the second you want to change something without upsetting the rest: a color, an element, a shot. If a model does both well, the workflow gets shorter. If it does not, you end up regenerating everything each time. The word "consistency" points in the same direction. In a short film, consistency means that people, objects and settings remain recognizable from one moment to the next. The source states that it is stronger, but provides no metrics. It should therefore be verified by looking at results on your own material: a logo that changes shape between one scene and the next is a concrete problem for anyone who sells. ## What is the 30% reduction really worth? A 30% reduction is a percentage, not a price. Google states it on the [launch page](https://blog.google/innovation-and-ai/technology/developers-tools/introducing-gemini-omni-1-2-flash/), but the material we have does not give the starting price on which to calculate it. Without that number, nobody can say how much a video will cost, or whether it will be cheaper than what you pay today. To turn the 30% into a figure you need three things: the current official rate, the unit in which the service is billed, and the number of attempts you make to get an acceptable result. This last item is often the most overlooked. If you have to regenerate several times, the real cost per usable video is higher than the cost per generation. To understand how the price lists of these services are read, the explanation of the [price per million tokens](https://giallostudio.it/en/blog/price-per-million-tokens-explained)is useful: videos are not measured the same way, but the logic of breaking cost down into comparable units is identical. And for thinking about the relationship between quality and spending, there is the guide to [Economic AI Models Compared](https://giallostudio.it/en/blog/cheap-ai-models-compared). | Appearance | What the source states | What is missing | |---|---|---| | Clip length | Native video generation and editing with 15-second clips, according to the [source](https://blog.google/innovation-and-ai/technology/developers-tools/introducing-gemini-omni-1-2-flash/) | Formats, resolutions and usage limits | | Consistency | Described as stronger, according to the [source](https://blog.google/innovation-and-ai/technology/developers-tools/introducing-gemini-omni-1-2-flash/) | Verifiable measurements and examples | | Price | 30% lower, according to the [source](https://blog.google/innovation-and-ai/technology/developers-tools/introducing-gemini-omni-1-2-flash/) | Starting price and rates | | Availability | Not stated in the material | Date, geographic areas, plans |

Worth remembering. A discount percentage says that the price goes down, not where it ends up. Before setting a budget, ask the provider for the official rate and calculate the cost per truly usable video, not per single generation.

## Which Italian SMEs and professionals can benefit? The examples that follow are hypothetical: they are meant for reasoning, and do not describe real clients or results achieved. **A small farm with direct sales.** It has seasonal products and little time to produce content. A short clip, the kind suited to a social media story or a product page, may be enough to present a new package format. The possible advantage is starting from a draft instead of from filming. The risk is that the product in the video does not look like the real one: for a food product, a faithful image is a matter of fairness toward the customer. **An architecture or interior design studio.** It often has to communicate an idea to a client before having a final design. A short film showing a room, which can be modified by changing one element at a time, makes the discussion quicker. The editing function, if it maintains the stated consistency, is the one that matters most here. The limit is obvious: it is an illustration, not a technical document, and it must be presented as such. In both cases the same rule applies: the video is an aid to communication, never proof of what you sell or build. To choose the right tool for each job, you can look at the overview of the [best AI models for video](https://giallostudio.it/en/blog/best-ai-video-generators), which compares the options by use. ## What mistakes should you avoid before using it in your company? There are four, and they almost always stem from haste. ### Doing the math on the 30% without the starting price It happens because the percentage is the only number available and seems sufficient. You avoid it by asking for the official rate and building the comparison with current spending from there. ### Judging consistency from a single successful example A well-made result is easy to find, and gladly shown. You avoid it by trying the same type of scene several times with your own material and noting how many tries are needed for an acceptable result. ### Publishing without human review A generated video can contain wrong details that go unnoticed at first glance: a garbled caption, an object that changes. You avoid it by assigning a person the task of watching every clip before publication, as you would for any other company content. ### Forgetting rights and transparency Whoever publishes a video created with AI must ask who holds the rights and whether its origin needs to be disclosed. The rules on this point must be checked case by case: the guide on [copyright and AI-generated content](https://giallostudio.it/en/blog/ai-generated-content-copyright) is a good starting point, and the one on how to [verify AI images, audio and video](https://giallostudio.it/en/blog/synthid-detector-verify-ai-content) shows what exists to recognize synthetic content. ## How do you apply it in your company, step by step? The safest approach is a small test, with a clear goal and a person in charge.
  1. Choose a single use case. For example, a short clip to present a product or a service, not your entire video communication.
  2. Gather the official rates. Check the provider's website for prices, plans and availability in your area: the material we rely on does not report them.
  3. Write the request with care. A precise description of scene, style and what must not appear reduces attempts. The guide to prompts for creating videos with AI contains practical examples.
  4. Record the attempts. Note how many generations are needed for a usable result and how much time the review takes.
  5. Have the result approved. A person watches the clip, checks fidelity to the product and rights, and decides whether to publish it.
  6. Compare with current spending. Only at this point, with the real numbers from your tests, does it make sense to decide whether to continue.
If the experiment works, the next step is to build it into a repeatable workflow, with written rules on who can use it and for what. Those who already have an AI use policy can update it; those who do not can follow the guide on how to write a [corporate policy on the use of AI](https://giallostudio.it/en/blog/company-ai-use-policy). ## How do you choose between this model and the others? There is no best model in absolute terms: it depends on the job, the budget and the level of control required. For a short presentation video, consistency may matter most; for intensive use, the cost per usable clip will matter; for use with confidential data, the provider's contractual terms will matter. A useful approach is to start from the result you need and work backward. The guide on [how to choose an AI model](https://giallostudio.it/en/blog/how-to-choose-ai-model) proposes a method that also applies to video models, without having to chase every announcement. ## What can a company concretely do with it? Based on the available material alone, a company can do one sensible thing: evaluate Gemini Omni 1.2 Flash as a tool for drafts and short clips, comparing the official rates with what it spends today on the same work. It cannot yet, however, plan a final budget or take availability in its own area for granted. If you want to understand where generative AI can truly serve in your processes, and where it does not pay off, you can start with an [AI consultation](https://giallostudio.it/en/ai-consulting): a short test is set up, with clear success criteria, before any investment. ## Frequently asked questions ### What is Gemini Omni 1.2 Flash? It is a Google model presented on the company's official blog. According to the material, it offers native video generation and editing, greater consistency and a lower price. Technical detail is not reported, so for complete data you should consult the [official page](https://blog.google/innovation-and-ai/technology/developers-tools/introducing-gemini-omni-1-2-flash/). ### How long are the videos it can create? The source mentions native video generation and editing with 15-second clips, as reported on [Google's page](https://blog.google/innovation-and-ai/technology/developers-tools/introducing-gemini-omni-1-2-flash/). The material gives no further limits on formats or resolutions. These should be checked in the official documentation. ### How much does it cost to use? Google states a price that is 30% lower, as the [source](https://blog.google/innovation-and-ai/technology/developers-tools/introducing-gemini-omni-1-2-flash/) indicates. However, the material does not report the starting price, so the cost of a video cannot be calculated. You need the provider's official rates. ### Is it already available in Italy? The material we consulted does not state the general availability date or the geographic areas. We therefore cannot confirm it. Before planning any use, check access on Google's official website. ### Can I publish the videos it creates without checking them? We advise against it. A generated video may contain errors of detail or may not faithfully represent a product. A person should always review the result and check rights and transparency obligations before publication. ### Is it worth replacing the video tools I already use? There is no answer that fits everyone. The comparison should be made using official rates and a test on your own material, also counting the attempts needed to get a usable result. If your current tool works and costs little, there is no rush to switch.