Grok Imagine Image 2.0: Quality Mode live on web and apps

xAI says Imagine Image 2.0 is generally available as the new Quality Mode on grok.com/imagine and on the Grok iOS and Android apps (xAI product note). The Aug 7 note frames 2.0 as image generation and editing built for real creative work: tighter instruction following, sharper small text and multi-part layouts, and edits that preserve inputs across generations.

Access is split. The Grok account says any signed-in user gets free limited tries; paid plans draw from a shared weekly compute pool across Imagine, Chat, and other features, with remaining balance and reset time under Settings → Usage, and extra credits starting at $5 (Grok on X, Aug 10 usage). API access for Imagine Image 2.0 is coming soon with no date published; prior Imagine models remain on the existing API (Grok on X, Aug 10 API).

The product note lists the editing kit that defines the SKU. Magic wand targets a region and leaves the rest alone. Segmentation picks precise areas. Background removal exports a subject with a transparent background. Multi-ref editing takes up to five input images in one generation. Smart resize fills a chosen aspect ratio from a single image. xAI also ships task templates for photo edit, product color change, headshots, icons, game assets, and similar starting points. Video stays on a separate line (Imagine Video 1.5); this card is the stills Quality Mode cut.

On the Arena tables xAI cites as of Aug 7 (models listed under SpaceXAI), Image 2.0 sits second overall on both the image-edit and text-to-image leaderboards in that snapshot (product note). Those ranks are vendor-cited Arena Elo figures from the note, not an independent Beam42 re-score of every production workflow.

Through Aug 9-10 the Grok account kept restating the consumer path: Image 2.0 is live as Quality Mode at grok.com/imagine, with precision region edits, sharper text, multi-ref compositing, and smart resize (Grok on X, Aug 9). Elon amplified the product on Aug 10. He said Grok Imagine is focused on professional usefulness, fun for consumers, and overall ease of use under a Design Arena daily-usage chart for Grok 4.5 apps (Elon on X, morning). Later he replied "Grok ftw" to a side-by-side topology prompt test that credited Imagine 2.0 on a genus-1 coffee-mug render (Elon on X, evening). That comparison is one social demo, not a published xAI topology benchmark.

Confirmed

• xAI product note (Aug 7): Imagine Image 2.0 generally available as Quality Mode on grok.com/imagine and Grok iOS/Android; editing tools (magic wand, segmentation, background removal, multi-ref up to 5, smart resize); templates; Arena image-edit and text-to-image ranks cited second overall as of Aug 7; API access coming soon (product note)
• Grok: Image 2.0 live as Quality Mode with precision edits, multi-ref, and smart resize (Grok Aug 9)
• Grok: free limited tries for signed-in accounts; paid shared weekly pool; Settings → Usage; extra credits from $5 (Grok Aug 10 usage)
• Grok: Imagine Image 2.0 API coming soon, no date; prior Imagine models already on API (Grok Aug 10 API)
• Elon: Imagine focus on professional usefulness, consumer fun, ease of use (Elon morning); "Grok ftw" on Imagine 2.0 topology comparison (Elon evening)

Not confirmed

• Exact free-tier try counts or paid weekly pool sizes by plan
• Imagine Image 2.0 API GA date or pricing
• Whether Speed Mode on the same surface still routes to the prior stills model in every region and app build
• Independent re-runs of Arena ranks after Aug 7
• Parameter count, training data, or open-weight status for Image 2.0
• That every side-by-side social demo used identical Quality Mode settings

What to watch

• Imagine Image 2.0 API launch note and docs model id
• Arena reruns and any xAI update to the Aug 7 Elo snapshot
• In-app strings that name Quality Mode vs Speed Mode defaults
• Template and editing-tool changes on grok.com/imagine

Primary posts: Elon evening · Grok usage · Grok API · product note