Fuel your creativity with new generative media models and tools

This Google DeepMind article centers on the release of a set of generative media models and tools aimed at creators. The core updates include the video model Veo 3, the image model Imagen 4, the AI filmmaking tool Flow, and an expanded rollout of the music model Lyria 2. Veo 3 is the most eye-catching update; Google says it not only improves on Veo 2's video quality but, for the first time, can add audio when generating video—for example, traffic sounds of a city street, birdsong in a park, and even dialogue between characters. It also emphasizes understanding of text and image prompts, a sense of real-world physics, and lip-sync capabilities. In the U.S., Veo 3 was made available on launch day to Ultra subscribers of the Gemini app and to Flow, and is also open to enterprises via Vertex AI.