Google DeepMind released Gemini Omni 1.1 Flash on Aug. 27, 2026, adding scene extensions, keyframe controls, and 4K upscaling to its generative video model. Developers can access the production-ready system through the Gemini API in Google AI Studio and the Gemini Enterprise Agent Platform.
Video creators can lengthen scenes in 10-second increments up to a total run time of 40 seconds. Omni 1.1 evaluates up to 10 seconds of prior context to preserve narrative continuity, an increase from earlier versions that referenced only the final second of footage.
Developers can also anchor the beginning and end of a sequence by specifying first and last frames. The model builds continuous video between those keyframes, facilitating zoom transitions, camera orbits, and seamless loops. For rapid testing, the platform generates 360p resolution previews at one-third the cost and up to 60 percent faster than standard 720p output, based on system throughput, before users upscale the footage to 1080p or 4K.
Partners including Adobe, Figma, and Runway have incorporated the system into their creative tools. Adobe added Omni Flash to Adobe Firefly, while Figma embedded it into Figma Weave to support multi-version branching. GMI Cloud also deployed the model for clients building educational media.
Subscribers to Google AI Plus, Pro, and Ultra tiers can use Omni 1.1 globally in Google Flow starting immediately. Scene extension is also accessible to those same subscription tiers within the Gemini application, alongside enterprise access via Google AI Studio and the Agent Platform API.
