Google launches Gemini Omni 1.2 Flash for cheaper cinematic AI video

Breaking the 10-second barrier

Google officially released Gemini Omni 1.2 Flash on November 5, 2026, to streamline the creation of high-quality AI video. Previously, creators often had to string together several short 5- or 10-second segments to build a longer scene. This process frequently caused visual drift, where the appearance of objects or characters changed unnaturally between cuts.

The new model generates native 15-second clips in a single pass, providing a smoother experience for narrative storytelling. It also offers smarter granularity, allowing users to direct the model in exact 1-second increments. This flexibility helps creators fit specific requirements for social media ads or explainer videos without manual stitching.

As nokiapoweruser.com reported, these improvements are designed to solve the pain points of rigid prompts and clipped sequences. By providing longer, continuous motion, the model supports complex camera movements like FPV drone passes and cinematic orbits. These features aim to provide professional-grade creative control to a wider audience of developers and businesses.

Lowering the cost of production

One of the most significant updates in the Gemini Omni 1.2 Flash release is the reduction in production costs. The API pricing has dropped by 30%, reaching approximately $0.07 per second for 720p footage. This translates to a cost of $12 per million video tokens, making it easier for teams to scale their content output.

To further manage budgets, the model supports a drafting and upscaling workflow. Creators can render low-resolution drafts at 360p or 720p to test their creative ideas quickly and cheaply. Once a specific shot is finalized, they can then upscale the winning render to high-definition 1080p or 4K output.

This pricing structure is intended to help businesses craft multiple commercial ad variants without blowing their budgets. By lowering the technical and financial barriers, Google is positioning this model as a tool for high-volume content production. It allows for more experimentation and iteration during the early stages of a project.

Conversational editing tools

Standard AI video tools often suffer from prompt roulette, where changing one small detail forces the user to regenerate the entire video. Gemini Omni 1.2 Flash introduces stateful multi-turn editing to solve this problem. This system allows users to chat directly with the model to make specific adjustments turn-by-turn.

For example, a user can ask the model to change the background lighting to a golden hour sunset or replace a blue mug with a ceramic coffee cup. Because the system tracks interaction history using unique identifiers, it only modifies the requested elements. This keeps the character's facial expressions, typography, and camera movements steady across different iterations.

This conversational approach removes the need for users to write complex master prompts from scratch every time they want a change. It allows for a more intuitive creative process where the AI understands the context of the previous turn. This level of control is a major step toward making AI video production feel more like traditional editing software.

Cinematic controls and consistency

The model also introduces advanced features for visual reliability and character consistency. It supports multi-speaker conditioning, which delivers natural speech and accurate lip-synchronization for dialogue. These updates have helped the model top evaluations on platforms like Arena.ai for text-to-video reliability.

For brand-heavy projects, the model can adhere to strict references. Users can feed up to 6 reference images, such as character design sheets, product angles, or brand colors, into the system. The model is designed to hold onto these fine details, ensuring that a character remains consistent from the first frame to the 360th frame.

These capabilities make it easier for professional creators to maintain a unified look across a series of clips. Whether it is a dramatic zoom-out or a sweeping orbit, the model aims to avoid sudden visual jumps or scale distortions. This focus on stability is critical for anyone trying to produce polished, cinematic content rather than experimental clips.

Availability and integration

Google has integrated Gemini Omni 1.2 Flash into several existing products and platforms. Users can access it through Google Flow for direct multi-turn scene creation and Google Vids for corporate presentations and marketing teasers. The Gemini App also includes the model for Google AI Ultra subscribers on both mobile and desktop devices.

For developers and third-party companies, the model is available via the Google Cloud and the Gemini API. This allows platforms like InVideo, Picsart, and Photoroom to embed these video generation powers directly into their own creative suites. This wide distribution ensures that the technology reaches a broad range of creative professionals.

By blending affordability with fine-tuned conversational editing, Google aims to make storytelling more effortless. The goal is to shift the focus away from wrestling with complex prompts and toward crafting compelling stories. As the tool becomes more accessible, the barrier for high-grade AI video production continues to drop for creators worldwide.