Google, Gemini Omni 1.1 Flash By publishing the model, video production with artificial intelligence is achieved with scene extension, first and last frame control. 4K upscaling added options. The model renders the scene using the last 10 seconds of the current video as context. Up to 40 seconds total in 10 second steps can sustain. In addition, it becomes possible to produce faster drafts at 360p resolution. Google makes the model available via Gemini API, Google AI Studio, Gemini app and Google Flow.
With Gemini Omni 1.1 Flash, Google is expanding the video tools of the previous Gemini Omni Flash preview release. While the previous model focused on high-speed video production and natural language video editing features, the new version directly video extension brings support. While extending a clip, the model examines the last 10 seconds of the original video and tries to preserve the image, movement, characters and sound in the following section. However, the system can make changes to the last frames of the input video to make the transition seamless. Each extension adds 10 seconds to the video and the total duration can reach 40 seconds.
The new model is in a single video output created 3 to 10 seconds It offers a duration option between and can watch videos per second. 24 squares It prepares at speed. Gemini Omni 1.1 Flash accepts text, images and video on the input side, while producing video on the output side. The duration of the source video uploaded for editing or extension can reach a maximum of 10 seconds, but different rules apply when repeatedly extending the videos previously produced by the model. Additionally, video stretching only adds new footage to the end of the clip, not to the beginning or middle of the existing video.
First and last frame control It allows defining two images as the starting and ending points of a shot. Gemini Omni 1.1 Flash produces video between these frames, maintaining camera movement and transition between two designated images. Thus, two separate frames can be used for camera turns, zooms, or videos whose beginning and end are connected. It is also possible to add images and videos as references. According to Google’s developer document, the system accepts a maximum of three reference videos, and each of these videos up to 3 seconds It can extend.
Gemini Omni 1.1 speeds up Flash video production
Google also eliminates the need to prepare each attempt in high resolution when producing videos. Gemini Omni 1.1 Flash alongside standard 720p option Draft video in 360p resolution can create. According to the company, 360p mode delivers results up to 60 percent faster than standard 720p production and cuts costs by a third. This allows you to try different scenes and compositions at low resolution and then move the preferred version to high resolution. The Google Flow mobile app also uses this option to prepare quick video drafts on the phone.
Resolution options 360p, 720p, 1080p and 4K The system creates 720p video by default. In addition, Google applies upscaling in 1080p and 4K options instead of producing directly at this resolution. Therefore, after low-resolution trials, it is possible to move the selected video to a higher resolution. On the Google Flow side, the company positions 1080p and 4K export options for high-resolution digital content, social media and broadcast editing processes. Gemini API documentation also clearly defines 1080p and 4K output as “upscaled” video.
Gemini Omni 1.1 Flash can process text, images and video together during video production and follows editing commands via natural language. Thus, it is possible to describe only the part that needs to be changed in the created video and request that other elements be preserved. The model also combines knowledge of physics with Gemini’s knowledge of history, science, and cultural context. Google also creates an audio track suitable for the video upon request, but the current API version does not support loading audio files as a reference. What the viewer cannot see in all Gemini Omni videos produced SynthID digital watermark is also included.
Google for developers Gemini Omni 1.1 Flash model gemini-omni-1.1-flash It offers it as a stable version with its code. The model can be used in Google AI Studio via the Gemini API, and corporate developers can access the model via the Gemini Enterprise Agent Platform. In addition, Google AI Plus, Pro and Ultra subscribers can use Gemini Omni 1.1 Flash features worldwide within Google Flow. The stage extension feature is also available to Gemini application users with the same subscriptions. Thus, Google is expanding the Gemini Omni distribution, which it launched in May with the Gemini application, Flow and YouTube Shorts, to developer tools.
Gemini Omni Flash was first introduced during Google I/O in May 2026, and the company announced the model as a generative media model that can process text, images, audio and video together. Google later on June 30 gemini-omni-flash-preview made its model available for public preview via the Gemini API. That version created 3- to 10-second videos at 720p resolution and allowed users to edit the videos with spoken commands. company now Gemini Omni 1.1 Flash completed this preview process and made the model available for general use.
According to Google’s Gemini API release notes it is outdated gemini-omni-flash-preview end point On September 30, 2026 will reach the end of its useful life. Therefore, developers using the preview model in question gemini-omni-1.1-flash Needs to upgrade to version. With the transition, the new model adds video stretching, video creation between the first and last frames, and resolution control to the API side. Google also provides video references and examples of scene stretching and resolution upscaling in its developer documentation and prompt guides.
Join Channel