Google introduces Gemini Omni 1.1 Flash for video creation and editing via API

On August 27, 2026, Google announced Gemini Omni 1.1 Flash, a model for video generation and editing identified as gemini-omni-1.1-flash. It is available to developers through the Gemini API in Google AI Studio. Google also lists its deployment in the Gemini Enterprise Agent Platform.

According to the technical documentation, the model accepts both text and image inputs. Through the Interactions API, it can create video and then edit it iteratively, meaning it can make further changes to an already generated output.

Gemini Omni 1.1 Flash brings longer scenes and controlled transitions

Announced features include extending an existing scene in 10-second increments up to a total length of 40 seconds. The model also supports interpolation between the first and last frames, which is intended to enable a transition between two specified images.

Another option is video references up to three seconds long. Google lists them as input material for the model alongside text and images. The documentation also confirms support for subsequent, conversationally requested video edits.

Outputs can be generated at 360p and the default 720p resolution. Upscaled 1080p or 4K should also be available. For 360p previews, Google claims speeds of up to 60% higher and one-third the cost compared with 720p. However, the announcement did not provide more detailed conditions or methodology for these measurements.

Availability through the API, Flow, and the Gemini app

According to Google, Gemini Omni 1.1 Flash is accessible through the Gemini API in Google AI Studio. Google also plans to deploy the model’s features in Google Flow, where they should be globally available to subscribers of the Google AI Plus, Pro, and Ultra plans.

Scene extension should also be available for these subscription tiers in the Gemini app. However, the announcement does not detail the full availability of all features by country, or the specific limits for individual resolutions, input materials, or editing uploaded video.

For developers, the combination of generation and editing in the API is particularly important. Instead of creating a short clip just once, they can incorporate scene extension, specifying the frames at the edges of a transition, and further changes to the generated video into their workflow.

What still needs to be verified in practice

In its materials, Google uses claims about studio quality, professional readiness, and improved consistency. These are marketing claims by the manufacturer, not independent comparative tests. It has not yet been independently confirmed how reliably the model will preserve character identity, scene continuity, or the accuracy of requested edits in real production workflows.

API pricing, the specific limitations of individual modes, and regional availability will also be important for deployment. The rules for labeling AI-generated videos, copyright protection, and restrictions on working with people’s likenesses will also require attention.

Upcoming tests will show how Gemini Omni 1.1 Flash compares in quality, consistency, and speed with competing generative video models.

Verified and updated: 08/27/2026 20:33