Google announced a new video model, Veo 3.1, with improved audio output, fine-grained editing controls, and improvements to image-to-video output. Veo 3.1 is based on Veo 3, which was released in May, and is said to produce more realistic clips and improve the ability to follow prompts.
Google says this model allows users to add objects to their videos and blend them into the clip’s style. Users will soon be able to remove existing objects from videos in Flow.
Veo 3 already has editing features such as adding reference images to animate characters, providing first and last frames to generate clips using AI, and the ability to enhance existing video based on the last few frames. In Veo 3.1, Google is adding audio to all of these features to bring your clips to life.
The company is deploying this model in its video editor Flow, Gemini app, along with Vertex and Gemini APIs. Since Flow was launched in May, users have created more than 275 million videos on the app.
