Sulat.com
AI models
$10 off the fastest DeepSeek V4.1 Flash, Kimi K3 and GLM 5.3 from Synthetic
ZenMux logo

Model details

Grok Imagine Image 2.0

Grok Imagine Image 2.0 is positioned by xAI as a high-quality image model aimed at production-ready creative assets rather than one-off novelty renders. The release emphasizes planning typography and layout, following detailed prompts, and preserving supplied elements across edits, with coverage of photography, graphic design, and illustration in a single workflow. Consumer users reach the model through the new Quality Mode inside Grok Imagine on the web, iOS, and Android, where tools such as a magic wand for localized edits, region selection, transparent background removal, smart resize, and goal-oriented templates for product shots, headshots, collages, icons, game assets, and merchandise make it approachable for non-developers as well as technical teams.

Developers access the same named model, grok-imagine-image-2.0, through the Imagine API, which returns images at 1K or 2K resolution across Low and Medium quality tiers and is served from US East and US West regions with a baseline rate limit of six requests per second that scales with account tier. Its distinguishing capabilities are region-level editing that replaces whole-image guessing, the ability to combine up to five reference images for multi-reference composition, and a smart-resize function that automatically fits nine common aspect ratios. Together these traits suit teams that need dependable typography, layout control, and localized corrections in marketing, e-commerce, and design pipelines where preserving specific elements across revisions is critical.

ZenMuxx-ai/grok-imagine-image-2.0grok

Quick Info

Powered by
Provider
ZenMux
Model key
x-ai/grok-imagine-image-2.0
Release date
Aug 7, 2026
Last updated
Aug 7, 2026
Input modalities
Output modalities
Capabilities

Limits

Output tokens
0 tokens
Context window
66,000 tokens

Latest news about Grok Imagine Image 2.0

ZenMux

Official sourceAnnouncement

xAI announced Grok Imagine Image 2.0 on August 7, 2026, as the new Quality Mode in Grok Imagine on grok.com and the iOS and Android apps. The model is positioned around a single goal: producing images usable in real work, with instruction-following that extends to fine details. Image 2.0 is trained for fidelity across Image 2.0 introduces precise editing tools: a magic wand that edits only the region you point at, segmentation for selecting precise image areas, and background removal that exports subjects on transparent backgrounds. Multi-reference editing accepts up to five input images in a single generation, removing the need for

Videos about Grok Imagine Image 2.0

More models around Grok Imagine Image 2.0