Grok Imagine Image 2.0
Explore xAI's Imagine Image 2.0 release and official X examples, then create, edit, segment, or refine selected areas with the Image 2.0 models integrated through KIE.
Model access: Grok Video sends Image 2.0 tasks through KIE's exact text-to-image, image-edit, segment-map, and segment-edit models. xAI's release page and Grok's X thread remain the sources for product highlights and examples.
Review two visual demonstrations from Grok's official Image 2.0 release thread, with each original X source and disclosed limitation kept visible.
These posts remain hosted by X. Creator attribution and the original source stay attached to every embed.
Grok demonstrates Magic Wand, Segmentation, and Smart Resize as the release's targeted editing workflow.
Official Image 2.0 release example; the exact prompt and generation settings were not disclosed.
The public X post will load when this section approaches the viewport.
The official thread presents infographics, ads, game assets, UI and UX mockups, and storyboards as Image 2.0 use cases.
Official release reel; the post says API access is coming soon and does not disclose exact prompts.
The public X post will load when this section approaches the viewport.
Choose Create, Edit with references, or Select areas, then provide the inputs required by that Image 2.0 workflow.
Use Create for text-to-image, Edit with references for a guided change, or Select areas to map and edit parts of an owned image.
Write a focused prompt for creation or selected-area editing. Reference editing accepts one to five images and also allows an optional prompt.
Create supports 1:1, 2:3, 3:2, 16:9, and 9:16. Reference editing adds Auto, and the current credit cost is shown before submission.
Submit the task and download the stored result. In Select areas, first create a segment map, choose returned areas, then describe the targeted edit.
xAI's release focuses on instruction following, typography and layout, precise editing, reusable creative tools, and practical visual production.
Magic Wand, segmentation, background removal, targeted edits, and up to five reference images support controlled changes while preserving the wider composition.
xAI highlights stronger text rendering and layout control for ads, infographics, interfaces, storyboards, and other structured visual assets.
Smart Resize and templates adapt finished work across formats, while consistent worlds support related assets for campaigns, games, and product concepts.
xAI reports that Image 2.0 ranked second in both text-to-image and image-editing Arena evaluations as of August 7, 2026.
The generator maps each visible workflow to the corresponding Grok Imagine Image 2.0 model available through KIE.
Create a new image from a written prompt and choose one of five supported aspect ratios.
Edit with an optional prompt, Auto or a fixed ratio, and one to five JPEG, PNG, or WebP references up to 30MB each.
Upload a local image owned by your account to identify selectable areas before editing.
Choose indexes returned by the segment map, write the requested change, and preserve the parent-child relationship in private history.
Grok Video currently accesses four Grok Imagine Image 2.0 models through KIE. This provider integration is separate from a claim of direct xAI API access.
New creation tasks use grok-imagine-image-2-0/text-to-image.
Reference edits use grok-imagine-image-2-0/image-edit with one to five verified image URLs.
Area detection uses grok-imagine-image-2-0/segment-map with one verified provider task or owned image source.
Targeted changes use grok-imagine-image-2-0/segment-edit with the segment-map task, a prompt, and validated mask indexes.
Answers about the release, KIE model access, four generation capabilities, X examples, uploads, and downloads.
Create from text, edit with references, or select and refine specific areas with the Image 2.0 models available through KIE.
