Gemini Omni 1.2 Flash: Google's AI video model makes 15-second clips, 30% cheaper

Google released Gemini Omni 1.2 Flash, a new version of its AI model that generates and edits video, and made it generally available on Thursday, 5 November 2026, the company said in a blog post. The model can produce video clips of up to 15 seconds, up from 10 seconds in the previous version, Omni 1.1, and costs $0.07 per second of 720p video through the Gemini API, which reports describe as about 30% cheaper than before. It is available to developers through the Gemini API and Google Cloud, and to users of Google Vids and Gemini for business, the company said. From Thursday, 3 December 2026, requests sent to Omni 1.1 will automatically be routed to 1.2 at the lower price.

Generative video, where an AI system turns a text description or an image into a moving clip, has been one of the most expensive kinds of AI to run. Each second of video requires far more computing than a paragraph of text. Price per second is therefore the number that decides who can afford to use it: an advertising agency making a few clips, or an app that wants to generate thousands for its users. A lower price and longer clips widen the range of uses, from short product videos to social media content and training material.

Google describes the model as fast and cost-effective, which points to high-volume use rather than the highest-end output. "Generally available" means Google is offering it with its normal service commitments, so businesses can build products on it rather than only experiment. These are the company's own descriptions; independent tests of quality and speed were not part of the reports.

For India, the change matters to the country's large community of developers, media start-ups and marketing agencies that build on cloud AI services. Indian companies making regional-language ads, explainer videos or e-commerce clips pay for these tools by the second, so a lower rate directly affects their costs. Through Google Vids and Gemini for business, the capability also reaches office users, not only developers. At the same time, cheaper and longer AI video makes it easier to produce convincing fake clips, which is one reason India's IT ministry has been pressing platforms to label AI-generated content.

Peepals calculation: If $0.07 a second is about 30% below the earlier price, the previous rate was roughly $0.10 a second (0.07 ÷ 0.70 = 0.10). At the new price, a full 15-second 720p clip costs about $1.05 (15 × 0.07), close to what a 10-second clip cost before (10 × 0.10 = $1.00). In other words, developers get about 50% more video for roughly the same money per clip. Our calculation from the reported figures.

Short-term: Developers using Omni 1.1 will be moved to 1.2 on 3 December 2026 and should test their applications before then. Businesses using Google Vids may see the new capabilities as the rollout reaches their accounts.

Long-term: Falling prices for AI video are likely to bring generated clips into more everyday products, from ads to education. That will raise questions about labelling, consent and copyright that regulators, including India's, are already working on.

Who is affected: app developers and start-ups using the Gemini API; marketing, media and e-learning companies; Google Workspace users with access to Vids; and audiences who will see more AI-generated video online.

What to watch next: the automatic move of Omni 1.1 API traffic to 1.2 on 3 December 2026.

Questions and Answers