Grok Imagine Image 2.0

Grok Imagine Image 2.0 is the second-generation image model from xAI. It was released on 8 August 2026 and became the default Quality Mode on grok.com/imagine and in the Grok apps for iPhone and Android. As well as making images from a prompt, it can edit them: you can select one part of a picture and change only that, remove a background, or combine up to five reference images, and keep refining in chat without starting over. xAI says it ranked second in the world on LMArena's leaderboards for both making and editing images, behind only OpenAI.

Who makes it

xAI is the AI company Elon Musk founded in 2023. Its main product is Grok, a chatbot built into X (formerly Twitter) and also available as its own app. That link to X matters, because images made with Grok can reach a huge audience within seconds.

History

When Grok first gained image generation in August 2024, it used FLUX from Black Forest Labs. In December 2024 xAI switched to Aurora, its own in-house image model. Grok Imagine launched in August 2025, first for paying SuperGrok subscribers, and added short six-second videos with sound alongside still images. Image 2.0 is the latest step, with a focus on editing and much better lettering.

Controversy

Grok's image tools have drawn more criticism than most. Imagine launched with a “spicy” mode that allowed sexual content, and within days journalists found it would make a semi-nude video of a Taylor Swift lookalike without being asked to. Around the turn of 2026 the problem grew much larger, as people used Grok on X to digitally undress photos of real women, and in some reports children, without their consent. Bloomberg estimated it was producing thousands of these images an hour. Regulators in several countries stepped in, and in the UK both Ofcom and the Information Commissioner's Office opened investigations.

The episode made Grok a test case for whether AI companies can be held responsible for what their image tools make.