Google on Thursday announced Gemini Omni 1.1 Flash, a production-ready update to its generative video model that adds scene extension, cinematic camera controls and 4K output, according to a post on the official Google blog by DeepMind product managers Anish Nangia and Alisa Fortin.
The update moves Omni from an experimental model into what Google calls production readiness for professional use via the Gemini API in Google AI Studio, with availability also listed through the Gemini Enterprise Agent Platform. Google describes Gemini Omni as having brought real-world reasoning to generative creation, and positions the 1.1 release as the point where the model becomes dependable enough for developers building video workflows, creative tools and media editing software.
Longer stories through scene extension
The headline capability is scene extension, which lets developers take an existing generated video and continue generating footage seamlessly from where it left off. With Omni 1.1, Google says the model can analyze up to 10 seconds of prior context — a leap from previous models that referenced only the final second. The result, according to the company, is improved visual consistency and narrative adherence when building longer stories or branching into new creative directions.
Videos can be extended in 10-second increments up to a total cumulative length of 40 seconds. That remains short of traditional video lengths, but for short-form content, advertising creative and pre-visualization work, Google is betting that control and consistency matter more than duration.
The demonstration material published alongside the announcement shows how the workflow is meant to function in practice: each new prompt continues the previous scene, with the model carrying visual context forward while the prompt directs changes in camera movement, setting and even mood — one sequence moves from a close conversation to a slow pull-out through dusty catacombs and then into a vast floating library. Building a scene in layers, rather than generating it in one shot, is closer to how an editor assembles footage than how early text-to-video tools worked.
Camera control and frame interpolation
The release also adds what Google describes as studio-quality video production tools. Among the examples shown on the announcement post: a cinematic dolly-zoom that moves the camera forward while zooming out, a mechanical snap-zoom into a character's eyes, and a 360-degree orbital rotation around a frozen character to showcase 3D depth and parallax.
Developers can also specify first and last frames for a shot, with the model interpolating smooth transitions between them — a technique familiar to professional animators that gives fine-grained control over where a shot begins and ends. Follow the latest AI developments as Google continues its rapid release cadence.
Iterate in 360p, deliver in 4K
Omni 1.1 Flash introduces a two-tier quality workflow aimed at cost control. Developers can generate quick 360p previews to iterate on ideas faster and more cheaply during the creative process, then upscale the final project to 4K resolution for a polished result. Google says the upscaling produces crisp, high-resolution output suitable for professional use.
The preview-to-4K pipeline addresses one of the persistent economic problems of generative video: the cost of regenerating full-resolution output every time a prompt changes. By decoupling exploration from delivery, Google is making the model more practical for commercial pipelines where dozens of iterations precede a final render.
Why it matters
The update reflects an industry shift from raw generation quality toward controllability — the ability to direct camera movement, maintain character consistency and hit exact frames, the way a director or editor would. For developers building creative software, the difference between a demo and a deployable product often comes down to these controls rather than raw fidelity.
Google's framing of the release makes the intended audience explicit. The company names three target categories — generative video workflows, creative tools and media editing software — and describes the updates as making generative video more controllable, faster to iterate on and polished for real-world deployment. Faster prototyping appears alongside 4K upscaling in the release summary, underscoring that iteration speed is being sold as a feature in its own right.
With Omni 1.1 Flash available in AI Studio and through the Gemini API, Google is also pushing the model toward enterprise users via the Gemini Enterprise Agent Platform, signaling that generative video is being positioned as business infrastructure rather than a consumer novelty.
What to watch
Pricing details for 4K output were not highlighted in the announcement, and developers will be watching how the 40-second cumulative limit affects real-world production plans. Competition in AI video remains intense, with rivals shipping their own controllability features at a rapid pace — a dynamic that has repeatedly shortened the shelf life of state-of-the-art claims this year.
For now, Google's message to developers is direct: generative video that once required prompt alchemy and luck can now be directed, extended and finished in 4K through a single API.
---
Stay Ahead of AIGet the latest AI news, analysis, and breakthroughs — all in one place.
Read more AI news →
