Google has introduced Gemini Omni 1.1 Flash, a generative AI model focused on video creation and editing. The model is being rolled out to developers and select subscribers, with features including video extension, frame-to-frame generation, reference clips and 4K output.
Video Extension Supports Up to 40 Seconds
Gemini Omni 1.1 Flash can analyze up to 10 seconds of footage from an existing video and extend the scene while maintaining visual and narrative continuity.
Developers can extend videos in 10-second segments, with support for video extensions of up to 40 seconds in total.
First and Last Frames Can Define Generated Footage
Developers can specify the first and last frames of a video, after which Gemini Omni 1.1 Flash generates the footage required between the two frames.
The feature supports complex camera movements such as camera orbits and zooms, as well as different transitions. It can also be used to create continuous and looping videos.
Reference Clips Support Video Generation
Gemini Omni 1.1 Flash supports the use of reference videos. Developers can use video clips of up to 3 seconds as references.
Reference clips can help retain a character's appearance, background and other visual details in newly generated videos. Google has also demonstrated the use of multiple reference clips to repeat different movements within the same video.
360p Mode Targets Faster, Lower-Cost Drafts
Google has added a 360p mode for creating faster and lower-cost video drafts. According to the company, 360p drafts can be generated up to 60 percent faster than 720p versions and cost about one-third as much.
The mode can be used for testing new ideas and creating storyboards. Once a project is finalized, developers can generate videos at 1080p or 4K resolution.
Availability Through Gemini API and Google Flow
Gemini Omni 1.1 Flash is being made available to developers through the Gemini API in Google AI Studio. Enterprise users can use the model through the Agent Platform API. Google has also released documentation, cookbooks and prompting guides for the new video tools.
Through Google Flow, the model is also being made available to AI Plus, Pro and Ultra subscribers. These users will also receive the scene extension feature in the Gemini app.
According to Google, platforms including Adobe, Figma Weave, GMI Cloud and Runway are integrating Gemini Omni Flash into their video workflows.
