Grok has unveiled Imagine Image 2.0, its next-generation image-generation and editing model aimed at creative and professional applications.
In an X update on August 8, 2026, the company said the model delivers greater precision, sharper text rendering, improved factuality and enhanced editing capabilities.
It allows users to make targeted changes, resize images and incorporate multiple reference images in a single generation while preserving key elements across successive edits. Grok said the model is designed to follow detailed instructions and produce visuals suitable for practical, real-world use.
“Imagine Image 2.0 is now generally available as the new Quality Mode on grok.com/imagine,” Grok stated, adding that the model is also available through its iOS and Android applications.
According to the company, Image 2.0 has been developed around the goal of producing images that are useful beyond experimentation, with the system designed to follow detailed instructions closely.
“2.0 follows instructions closely, down to the details. It plans typography and layout the way a designer would, so dense, multi-part visuals hold together and small text comes out sharp,” the company said.
Grok explained that the model also focuses on maintaining elements of an image when users make further changes, addressing the iterative nature of creative work.
“And it preserves what you put in across generations and edits,” it said.
A major feature of Image 2.0 is its precision editing capability, which allows users to make targeted changes without substantially altering other parts of an image.
“The magic wand edits the region you point at and leaves the rest untouched,” Grok said.
The model also includes segmentation tools for selecting specific areas of an image for modification. Another feature allows users to remove backgrounds and export subjects with transparent backgrounds, making them suitable for use in other creative projects.
Grok has also introduced multi-reference editing, which allows up to five input images to be used in a single generation.
“Multi-ref editing accepts up to 5 input images in a single generation, removing the need for manual compositing.”
The model further includes a smart resize feature intended to allow users to adapt an existing image to different aspect ratios.
“One image, any size: pick a ratio and the model fills in the frame,” Grok explained.
The company said Image 2.0 was trained to deliver fidelity across photography, design and illustration, while treating editing as a core capability rather than an additional feature.
Grok also elaborated that the model ranks second globally in both text-to-image generation and image editing.
The launch positions Imagine Image 2.0 as a tool aimed at users who require more control over AI-generated visuals, particularly when creating or repeatedly editing professional content.