Google launches Veo and Imagen 3, its new models for visual content creation

Google continues advancing in generative media innovation with the launch of Veo and Imagen 3, models designed to support creators. Veo, the most advanced model for generating high-definition videos, and Imagen 3, its most precise text-to-image model, offer a new way to create high-quality visual content.

Veo: The most advanced video generation model
Veo can generate videos in 1080p resolution, accurately representing a wide range of visual and cinematic styles. With an advanced understanding of natural language, Veo can capture the exact tone of a prompt and create videos that maintain coherence in long takes, where people, animals, and objects move realistically.

This model offers unprecedented creative control, interpreting cinematic terms like "timelapse" or "aerial shots" and creating content that flows naturally. Veo builds on years of research in generative video models, including previous works like Generative Query Network (GQN) and Imagen-Video.

Imagen 3: The highest-quality text-to-image model
Imagen 3 is Google's most advanced model for generating images from text. Its ability to produce photorealistic images with high detail and fewer visual artifacts makes it an ideal choice for creators. Imagen 3 better understands long prompts and translates them into images that capture small details, achieving an unprecedented level of precision.

Additionally, this model excels at creating text within images, something that has been challenging for other generation systems. This opens new opportunities for creating personalized content, from messages to presentations.

Google is making Veo and Imagen 3 available in preview, with plans to integrate these capabilities into platforms like YouTube Shorts in the near future.