Google's Imagen 3, a text-to-image model developed by DeepMind, is now reportedly available for public use through ImageFX. This model is known for its ability to generate high-quality images with enhanced detail, lighting, and prompt understanding. Initially, Imagen 3 is available to select creators as a private preview, but users have found ways to access it by selecting the US as their location during login, even without a VPN.
Imagen 3 is integrated into ImageFX, part of Google's AI Test Kitchen, which allows users to create images based on textual prompts.
The feature is designed to cater to a wide range of creative needs, from photorealistic landscapes to stylized art. It includes unique functionalities like seed control for image fine-tuning and prompt suggestions to aid creativity. Despite its capabilities, ImageFX has some limitations, such as generating only square images and lacking image editing features, which are areas for potential future enhancements.