Google’s latest text-to-image model, Imagen 3, is set to debut on the Vertex AI platform, offering select customers an advanced tool designed for faster and more accurate image generation. Initially previewed at Google I/O and available to select creators, Imagen 3 boasts enhanced photorealism, better understanding of detailed prompts, and superior text rendering within images. Senior Research Director Douglas Eck from Google DeepMind highlighted that Imagen 3 surpasses its predecessors in detail and accuracy, making it the most capable image generation model to date.
The model includes multi-language support, safety features like SynthID digital watermarking, and support for multiple aspect ratios. Companies like Shutterstock have already integrated Imagen into their systems, generating millions of images and appreciating its balance of speed and quality. However, Google has not disclosed when its Gemini AI will resume image generation due to previous inaccuracies. Google Cloud Chief Executive Thomas Kurian emphasized that Imagen and Gemini serve different purposes, with Gemini being a multimodal model capable of processing various types of inputs. Despite the advancements, questions about Gemini’s future functionality remain unanswered.
