Grok Imagine 2.0

• 🆕 xAI's newest Imagine image model, in its editing/inpainting configuration
• 🎯 Natural-language edits applied to an uploaded image or URL
• 🔧 Multi-image editing combines up to three source images per request
• 👁️ Handles subject combination, style transfer, and scene composition
• 🏢 Shipped as the "Quality Mode" experience across Grok's web and mobile apps
• 📚 Positioned by xAI for precise generation and editing in real creative work
• 💬 Same family as the text-to-image Grok Imagine 2.0 endpoint
• 🌐 Accessed through xAI's Imagine API alongside video and audio models

xAI

xAI is an American artificial intelligence company and wholly owned subsidiary of SpaceX. The company develops AI systems under the Grok brand, spanning language models, image generation, and video synthesis. xAI has quickly established itself as a multimodal…

Grok Imagine 2.0 (edit) is the image-editing side of xAI's latest Imagine image model, part of the same Grok Imagine 2.0 model family as its text-to-image counterpart Grok Imagine 2.0. Rather than generating from scratch, it takes an existing image plus an instruction and returns a modified version, covering inpainting-style local changes, restyling, and compositional edits. xAI describes the release as focused on precise image generation and editing built for real creative work, and made it generally available as the new Quality Mode on grok.com/imagine and the iOS and Android apps.

Within xAI's Imagine API, editing works by supplying a public image URL or a base64-encoded data URI together with a description of the desired change; multi-image editing accepts up to three source images in a single request so subjects can be merged, styles transferred, and scenes composed. Generation-side controls include output count, aspect ratio, resolution, and response format.

Compared with its same-family predecessors — the earlier Grok Imagine editor and the Grok Imagine High Quality editing tier, paired with Grok Imagine High Quality (SOTA) — the 2.0 generation is xAI's successor quality tier for editing. Detailed per-version capability deltas beyond that positioning are not published, so behaviour differences are best verified directly against xAI's Imagine documentation.