Veo 2 and Imagen 3.1: Google’s AI models transform video and image production

Google has announced the release of two new AI models, Veo 2 and Imagen 3 latest, aimed at enhancing video and image generation capabilities. These models are designed to cater to a broad range of users, from YouTube creators to enterprise clients, and are integrated into Google's creative tools like VideoFX and ImageFX.

Veo 2 is an advanced video generation model that produces high-quality videos with improved realism by understanding real-world physics and human expressions. It supports resolutions up to 4K and can create videos lasting several minutes. Veo 2 excels in following complex filmmaking instructions, such as specific camera shots and cinematic effects, making it a powerful tool for content creators. The model also addresses common AI generation issues like unwanted artifacts, producing fewer errors compared to previous versions.

Imagen 3 (3.1), the latest image generation model, offers enhanced color balance and detail, capable of rendering diverse art styles from photorealism to anime. It has been rolled out globally in ImageFX across more than 100 countries. Imagen 3 aims to provide users with more vibrant and accurately detailed images compared to its predecessors.

In addition to these models, Google introduced Whisk, a new tool that allows users to prompt with images to create unique visual content. Whisk combines Imagen 3 with Gemini's capabilities for visual understanding, enabling users to remix subjects and styles creatively.