Google has announced an updated version of its AI video model, Veo 3.1, which features improved audiovisual quality and a better understanding of prompts. The new version is being rolled out to Flow, the Gemini app, Vertex AI, and via Gemini APIs. Key new features include 'Ingredients to Video,' which allows users to provide multiple reference images to maintain consistent characters, objects, and style, and 'Frames to Video,' enabling the selection of starting and final scenes for seamless video generation. The model now also supports adding and, soon, removing objects within a video clip, as well as an 'Extend' option to create longer videos. For the first time, creators can also add generated audio to existing capabilities within the Flow platform.