Gemini Omni 1.1 Flash arrives with improvements to extend professional video clips up to 40 seconds

Google has released its model update Gemini Omni Flash with version 1.1, which offers enhanced video production by adding the ability to extend a scene up to 40 seconds, first and last frame interpolation, sharp 4K upscaling, and faster prototyping, among other new features.

The technology company unveiled its Gemini Omni generative artificial intelligence (AI) model within the framework of its annual developer event Google I/O 2026. Later, the Gemini Omni Flash version was officially launched, along with Nano Banana 2 Lite, as a development tool. multimodal video capable of generating 10-second videos with enterprise-level productivity at high speed.

Now, Google has released the update Gemini Omni 1.1 Flashwhich comes with improvements to make video production more controllable and faster to iterate, in addition to including capabilities such as function to extend a scene.

This is what the company explained in a statement on its blog, where it detailed that this function allows use an existing video and continue generating footage “without interruptions” from where it was interrupted.

For this, Omni 1.1 analyzes up to 10 seconds of previous context (instead of just analyzing the last second like the previous version) and, after that, generates More scene time with greater visual coherence and “better narrative adherence.” This allows users to create longer stories up to 40 seconds, expanding 10 second scenes into 10 seconds.

Another improvement applied to the model is the ability to generate fluid transitions and camera movements, interpolating the first and last frame of each shot. This is useful when generating “complex camera orbits, zoom transitions, or seamless loops,” as Google explained.

Omni 1.1 also generates Light previews in 360p resolution up to 60 percent faster and “at a third of the cost” compared to standard 720p resolution. Furthermore, it also allows video upscaling to 4K resolution, that is, it generates high resolution images at 1,080p or 4K for professional production.

Google has also specified that it allows add video references in the multimodal input to generate the content. Specifically, users can use as reference Up to three seconds of video to create a scene. As a result, the model will maintain visual context and character coherence that appear based on said references.

All in all, Omni 1.1 is rolling out to Google AI Studio, the Gemini Enterprise Agent platform API, and to all Google AI Plus, Pro, and Ultra subscribers globally on Google Flow. It should be noted that the function of extending scenes is also available to subscribers globally in the Gemini app.