Elon Musk’s Imagine Image 2.0, promoted on X, has become a viral hit for meme creation, ranking second in the world in both text-to-image generation and image editing on the large model arena. Unlike GPT-Image-2, released in April, Imagine Image 2.0 highlights interactive precise editing: after uploading an image, Grok automatically layers it, allowing users to select specific areas for modification. For example, a Wikipedia meme is segmented into layers including fragments of the Earth and the character holding it, divided into face and gesture parts. Users can click a download button to export any subject with a transparent background. A magic wand tool edits painted areas that automatic layering misses. The model also supports editing and merging multiple images, with multi-reference editing processing up to 5 input images at a time for automatic stitching. On the Imagine official website, users upload images, describe requirements, and are directed to a photo editing page. The right sidebar shows photo layers after automatic background removal, such as a red jacket, white T-shirt, and AI company logo. Users can click elements to modify them via prompts, relying on the model’s world knowledge. Precise editing allows framing objects not automatically recognized, and users can quickly edit existing memes, like the distracted boyfriend meme, by editing different selections.