Google reveals new AI models Veo and Imagen 3 to create lifelike videos, images with just text

Google reveals new AI models Veo and Imagen 3 that can create videos, images with just texts
Google has announced two powerful new AI models aimed at changing the way we create media. Unveiled at the Google I/O event, these models are designed to help creators produce high-quality videos and images more easily and accurately.

Veo for making videos
Veo is Google’s latest and most advanced video generation model. It can create high-definition 1080p videos that can last over a minute. Veo understands detailed natural language prompts, allowing it to accurately generate videos that match a user's creative vision. It can handle complex instructions like "timelapse" or "aerial shots of a landscape.”


During the Google I/O 2024 keynote, Google showcased its work with well-known creators, such as filmmaker Donald Glover and his studio, Gilga, to test and improve Veo. 

Imagen 3 for creating images
Imagen 3 is Google's newest text-to-image model, designed to produce realistic and detailed images from text descriptions. Google claims it offers better image quality with fewer visual errors and understands detailed prompts. Google claims this new AI model is more capable of creating lifelike images in various styles. It can also accurately generate text within images, which has been a shortfall most image generation models. Imagen 3 can be accessed via Google Labs. 

Music AI Sandbox
Google is also expanding its Music AI Sandbox with new tools that allow musicians to create and transform music using AI. The company has collaborated with artists like Wyclef Jean and Marc Rebillet to showcase the potential of these tools.