News in Short

• Google has introduced Gemini Omni 1.1 Flash with upgraded AI video generation and editing capabilities.
• The model can extend videos by up to 40 seconds, generate footage between two specified frames and use short video clips as references.
• It supports 4K video generation, while a new 360p mode offers faster and cheaper drafts.

Google has introduced Gemini Omni 1.1 Flash, an upgraded generative AI model focused on video creation and editing. The new model can extend existing footage, generate video between two specified frames and use short video clips as references. It also supports video generation at up to 4K resolution.

The update gives developers more control over AI-generated footage, while a new 360p mode is designed to make testing and rapid video prototyping faster and less expensive.

Gemini Omni 1.1 Flash Gets Longer Video Extensions

One of the biggest changes in Gemini Omni 1.1 Flash is its ability to understand more of an existing video when extending it.

The model can now analyse up to 10 seconds of preceding footage, compared with just the final second in earlier versions. Developers can add extensions in 10-second increments, with the total extension reaching 40 seconds.

This could help creators continue an existing sequence while maintaining greater consistency in elements such as the visuals and narrative.

The model can also generate footage between a specified beginning and ending frame. This allows users to define how a shot should start and finish while letting the AI create the transition between them. Google says the feature can handle movements such as camera orbits and zooms, as well as transitions for continuous or looping footage.

AI Model Can Use Video as a Reference

Gemini Omni 1.1 Flash also supports up to three seconds of video reference for a new generation.

Creators can use reference footage to help the model retain details such as a character’s appearance or the setting across generated footage. Google has also demonstrated using multiple reference clips to recreate different movements within a single continuous video.

For developers experimenting with ideas, Google is introducing a 360p generation mode. The company says 360p drafts can be generated up to 60 percent faster than 720p versions and cost one-third as much.

The lower-resolution option is aimed at tasks such as rapid prototyping and creating storyboards before generating the final footage at higher resolutions.

Once the sequence is finalised, developers can generate it at 1080p or 4K resolution.

Gemini Omni 1.1 Flash Availability

Developers can access Gemini Omni 1.1 Flash through the Gemini API in Google AI Studio, while enterprise users can use it through the Agent Platform API. Google has also released documentation, prompting guides and a cookbook for the new video-generation capabilities.

For consumers, the model is available globally to AI Plus, Pro and Ultra subscribers through Google Flow. Scene extension is also available to these subscribers through the Gemini app.

Google said the model is already being integrated into video workflows by companies including Adobe, Figma Weave, GMI Cloud and Runway.