On August 7, xAI fully launched Imagine Image 2.0. This version focuses on real-world work scenarios, supporting precise command following, layout processing, and maintaining original content. New features include area erasure, precise segmentation, and background removal, as well as support for multi-reference editing of up to five input images, significantly simplifying the iteration process. The model’s fidelity in photography, design, and illustration fields has been specially trained, and it ranks second in both text generation and image editing rankings worldwide. The platform provides a library of preset templates for photos, product images, and avatars, and offers the grok-imagine-image-2.0 API for developers to use.
On August 7, xAI launched Imagine Image 2.0, which was fully available on grok.com/imagine, iOS, and Android applications. Image 2.0 focuses on real-world working scenarios, can accurately follow instructions and handle layout and formatting, while preserving the original content from multiple edits. Its new features include area erasure, precise segmentation, background removal, and support for up to five input images per edit, significantly simplifying the iterative creation process. The model has been specifically trained for high fidelity in photography, design, and illustration, and ranked second in the world in both text generation and image editing categories. Additionally, the platform has introduced a library of preset templates covering scenarios such as photo editing, product images, and avatars, and made the grok-imagine-image-2.0 API available for developers to use.