Google has introduced a new version of its AI model for video creation — Veo 3.1, which is now available in the Flow editor, the Gemini app, as well as through Vertex and Gemini API. The model has improved audio creation, more detailed editing tools, and more accurately follows user prompts during video generation. Now in Flow, objects can be added to videos that organically fit into the style of the clip, and soon a feature for removing objects from the scene with automatic background and lighting restoration will be available.
Veo 3.1 expands the core functions of the editor, including “Ingredients to Video,” which allows creating scenes based on multiple reference images, and “Frames to Video,” where you can set the starting and ending frames and get a smooth video transition with audio accompaniment. The “Scene Extension” feature allows extending the last second of the clip by another minute, also with added sound. In all these functions, an audio track is now generated, making the videos more realistic.
The Flow editor has also received new capabilities for making changes to videos. With the “Insert” tool, users can add new elements to the scene, and the system automatically adjusts shadows and lighting for a natural look. The “Remove” feature will allow removing unnecessary objects, and the AI will restore the background so that the extraneous element is unnoticeable in the final video.
Veo 3.1 is available both for developers through the Gemini API and for business clients in Vertex AI, and anyone interested can use the model in the Gemini app. Various pricing plans are provided for users: a free plan with a limit of 100 AI credits per month, as well as paid subscriptions with a higher limit and access to the fastest version “Veo 3 Fast.”
