Google Unveils Veo AI Video Generator and Upgraded Imagen 3
Google introduces Veo, a powerful new AI model that generates high-definition videos over a minute long using cinematic prompts. The company also announces Imagen 3, an upgraded image generator that promises fewer artifacts and greater detail.
Google unveils Veo, a new artificial intelligence model designed specifically for video generation that understands both visual semantics and natural language. This advanced tool allows users to create high-definition 1080p videos that extend beyond a minute in length. Veo understands specific cinematic terms such as aerial shots and timelapses, enabling creators to tailor their video outputs to precise artistic styles.
Alongside Veo, Google introduces Imagen 3, which serves as the company's highest-quality text-to-image model to date. This third-generation upgrade significantly improves upon the Imagen 2 technology that currently powers the Gemini chatbot. Imagen 3 delivers a much higher level of detail in generated images while noticeably reducing visual artifacts and impurities.
Google builds Veo on years of previous generative video research, combining techniques from models like Phenaki, VideoPoet, and Lumiere to improve overall quality and resolution. The tech giant currently invites filmmakers and creators to test Veo so the model can learn to accommodate a wide variety of artistic use cases before a wider public release.