Google on Tuesday unveiled Veo 3, its latest AI video generator that sets itself apart by not only generating visuals but also incorporating audio—including character dialogue and ambient sounds like animals. This gives it a clear distinction from OpenAI’s Sora, which lacks audio generation.
According to Eli Collins, VP of product at Google DeepMind, Veo 3 performs well in lip-syncing, physics-based rendering, and interpreting both text and image prompts. It’s currently available to U.S. users through Google’s new Ultra plan, which costs $249.99 per month and is aimed at advanced AI users. Veo 3 is also accessible through Google’s Vertex AI for enterprise clients.
Alongside Veo 3, Google also introduced Imagen 4, an upgraded image-generation tool that delivers sharper, more realistic results from user prompts. Another major reveal was Flow, a filmmaking assistant that enables users to create cinematic content by simply describing locations, shots, and stylistic preferences. Flow is integrated across platforms like Gemini, Whisk, Workspace, and Vertex AI.
These releases reflect the growing demand for visual and multimedia capabilities in generative AI. OpenAI recently faced server strain from high demand for its GPT-4o image generator. While Google’s past image tools faced criticism for inaccuracies, the company has made improvements, including updates to Veo 2 and opening up Lyria 2, its music-generation model, to YouTube Shorts creators and enterprise users.
