Google has announced an updated version of their AI video model, Veo 3, with improved audiovisual quality and a better understanding of prompts. The company also introduced new capabilities like granular control for editing for the AI filmmaking tool, Flow, which is powered by the Veo model.

Users can give multiple reference images for what they want with ‘Ingredients to Video,’ to get similar characters, objects and style in Flow. They can also choose the starting and final scene with ‘Frames to Video,’ and generate a seamless video. 

The model also allows users to add objects to a video clip and blend it into the existing style, Google noted. Users will also be able to remove an object from a video soon.

There’s also an ‘Extend’ option so users can create longer videos, up to a minute or even more. 

Veo 3.1 will be rolled out to Flow, the Gemini app, Vertex AI and via Gemini APIs.