Google has launched Gemini Omni 1.2 Flash, an upgraded AI video-generation model designed to produce longer clips at a lower cost while maintaining better visual consistency. Announced on November 5, 2026, the model introduces native video generation of up to 15 seconds, compared with the 10-second limit of its predecessor, Gemini Omni 1.1 Flash. Google has also reduced the generation price by 30%, making the technology more accessible to developers, businesses and creative professionals.
The upgrade goes beyond generating longer videos. Gemini Omni 1.2 Flash introduces conversational video editing, improved reference-image accuracy, more natural dialogue and better character consistency across scenes. Google is positioning the model for commercial applications such as product advertisements, virtual try-ons, social media campaigns and video explainers, where businesses need to produce large volumes of content without repeatedly generating entire videos from scratch.
Gemini Omni 1.2 Flash Reduces Video Generation Costs by 30%
One of the biggest changes is the reduction in video-generation pricing.
Google has lowered the cost of generating 720p video to $0.07 per second, approximately 30% less than Gemini Omni 1.1 Flash. The reduction could make AI-generated video more practical for businesses that regularly produce advertisements, product demonstrations and promotional content.
The following table illustrates the API cost at the announced 720p rate:
| Video duration | Estimated generation cost |
|---|---|
| 3 seconds | $0.21 |
| 5 seconds | $0.35 |
| 10 seconds | $0.70 |
| 15 seconds | $1.05 |
| 60 seconds | $4.20 |
These figures are calculated using Google’s announced per-second price. They do not include other potential costs, such as additional editing, storage, application infrastructure or higher-resolution processing.
The lower price is particularly relevant for advertising agencies and e-commerce companies that may need hundreds or thousands of video variations. Instead of producing every advertisement through traditional filming and editing, businesses can use AI to generate initial versions and refine them through prompts.
AI Videos Can Now Run for Up to 15 Seconds
Gemini Omni 1.1 Flash supported video generation of up to 10 seconds. The new version increases the maximum duration to 15 seconds, providing additional room for storytelling and product demonstrations.
Users can specify durations in one-second increments between three and 15 seconds. This gives creators more flexibility when developing short advertisements, social media videos and visual explainers.
A 15-second generation can accommodate several visual developments within a single clip. For example, an online retailer could generate a product advertisement showing an item from multiple angles, followed by a lifestyle scene and a final product presentation.
The longer duration is also useful for character-based content, where facial expressions, movement and dialogue need more time to develop.
Google says the model can generate video at 360p and 720p for faster drafting, with selected outputs subsequently upscaled to 1080p or 4K. This workflow allows creators to experiment with ideas before spending resources on higher-resolution versions.
Better Character and Product Consistency
Maintaining visual consistency has been one of the major challenges in AI video generation.
Characters can change their facial appearance, clothing or proportions between scenes. Similarly, products may develop inconsistent packaging, colours or design details as the camera moves.
Gemini Omni 1.2 Flash is designed to improve reference adherence, allowing users to provide images, short video clips and audio references to guide generation.
Google says the model can preserve important details across different camera angles and scenes. This includes character appearances, clothing, product packaging, typography and brand colours.
For businesses, this improvement could make AI-generated advertising more practical. A retailer could provide product photographs and use them to generate multiple promotional videos while maintaining a consistent product appearance.
The model also supports reference-based workflows for virtual try-ons, product catalogues and animated demonstrations.
Conversational Editing Lets Users Modify Existing Videos
Another major addition is multi-turn conversational editing.
Instead of generating an entirely new video whenever a detail needs to change, users can refine an existing clip through follow-up instructions.
For example, a creator could generate a video featuring a product in a studio and then ask Gemini to replace the background with an outdoor environment. Further prompts could change the lighting, camera angle, weather or product presentation.
The model is designed to preserve the other elements of the original video while applying the requested modification.
This approach could reduce unnecessary generation costs and make the creative process more flexible.
The feature is particularly useful for advertising campaigns that require multiple versions of the same video. A company could create variations for different audiences, locations or marketing platforms without starting each project from the beginning.
More Natural Dialogue and Improved Audio
Google has also improved the model’s ability to generate speech and synchronise dialogue with characters.
Gemini Omni 1.2 Flash supports more natural-sounding conversations, clearer lip synchronisation and distinct voices for multiple speakers.
This could benefit businesses creating AI-generated explainers, animated advertisements and character-based videos.
For example, a short advertisement featuring two characters can include separate voices and more natural conversational timing. The model can also use audio references to guide speech generation, although Google says audio-reference access through the API is restricted to allowlisted users.
The improvements are intended to make generated videos feel more complete, rather than relying on silent visual sequences that require extensive post-production.
Google Expands Availability Across Its Products
Gemini Omni 1.2 Flash is now generally available through the Gemini API and Google Cloud, allowing developers and businesses to integrate video generation into their own applications.
The model is also being made available through several Google products and creative platforms.
| Platform | Availability and use |
|---|---|
| Gemini API | Developers can generate and edit videos |
| Google Cloud | Enterprise deployment and production workloads |
| Google Vids | Create clips up to 15 seconds |
| Google Flow | Video generation and conversational editing |
| Gemini app | 15-second video creation rolling out to Google AI Ultra subscribers |
| Google Ads | Video campaign workflows |
Google Vids users can access the updated model for generating clips up to 15 seconds. The rollout began on November 5 across eligible consumer, business and education plans.
The Gemini app rollout is more limited, with 15-second video creation initially being introduced for Google AI Ultra subscribers.
Google is also working with creative technology companies, including Adobe, Runway, Picsart, Krea, Artlist and InVideo, to bring the model into existing content-production workflows.
What Gemini Omni 1.2 Flash Means for the AI Video Industry
The launch highlights the increasing importance of affordability and controllability in generative video.
Earlier AI video systems attracted attention primarily because they could transform text prompts into realistic moving images. However, businesses need more than visually impressive demonstrations. They require consistent characters, accurate products, predictable editing and manageable production costs.
Google’s latest model attempts to address these requirements by combining lower pricing with longer generation times and conversational editing.
For small businesses, the technology could reduce the cost of producing promotional videos. E-commerce companies could turn product photographs into short advertisements, while marketing agencies could create multiple campaign variations more efficiently.
Nevertheless, AI video generation still has limitations. Google acknowledges that fast sports action, rapid script changes and complex physical interactions remain areas for further improvement. Businesses will also need to review generated content for visual inaccuracies, misleading representations and brand compliance.
The Bigger Picture
Gemini Omni 1.2 Flash strengthens Google’s position in an increasingly competitive AI video market. Companies such as OpenAI, Runway, Adobe and other generative media developers are working to make video creation faster, more realistic and easier to control.
The competition is gradually shifting from simply generating a convincing video to supporting an entire production workflow. Editing, reference accuracy, audio generation and cost per second are becoming important factors in deciding which model businesses adopt.
Google’s pricing reduction and 15-second generation limit could make its model attractive for high-volume commercial production, particularly in advertising and e-commerce.
Looking Ahead
The next challenge for Google will be improving reliability across complicated scenes while maintaining its cost advantage. Businesses will be looking for longer continuous sequences, accurate text rendering, stable character identities and more dependable physical interactions. Improvements in these areas could make AI-generated videos suitable for an even wider range of commercial applications.
For now, Gemini Omni 1.2 Flash gives developers and creators a more affordable way to produce and edit short AI videos. Its combination of longer clips, reference-based generation and conversational editing demonstrates how AI video tools are moving closer to becoming practical production platforms rather than experimental creative technologies.
Get the day’s top stories in your inbox
One concise email. No spam, unsubscribe anytime.
