Grok Image 2.0
Turn a written idea or reference photo into a finished visual with Grok Image 2.0. Generate, revise, restyle, and compare directions in one conversational workflow, with output up to 2K.
Overview
Grok Image 2.0, often searched for as Grok Imagine Image 2.0, is an AI image creation and editing experience that turns text prompts and reference images into new visuals. You describe the result you want, review the output, and continue refining it with plain-language instructions.
You can ask Grok Image 2.0 to change an object, adjust the lighting, apply a new visual style, or build a fresh composition from several references. Choose an aspect ratio that fits the intended placement, request multiple images when you need options, and produce images at up to 2K resolution.
Creative control
Grok Image 2.0 brings image generation and image editing into the same workflow. Each feature helps you move from an idea to a more specific, usable asset without starting over after every change.
Write what should appear in the frame, including the subject, setting, lighting, composition, and mood. Strong prompts name the main subject first, then add environment, camera view, palette, materials, and intended use.
Start with an existing image and describe the change you want. Restyle the source, add or modify details, and create a revised version with one focused instruction at a time.
Use the output of one edit as the input for the next. Establish the scene, adjust details, then refine the final style so each decision is easier to evaluate.
Provide two or more source images when a single reference does not contain everything you need. State which source defines the subject, setting, or style.
Restyle an image with a descriptive prompt, from realistic photography to anime, oil painting, pencil sketch, pop art, and watercolor, while keeping the core idea intact.
Choose an aspect ratio and resolution that match the destination, with output up to 2K. Request multiple images when comparing creative directions.
Examples
Workflow
Start with a prompt or a reference image, then improve the result through short, focused revisions in four clear steps.
Begin with a written concept when you need a new scene. Upload or select a reference image when you want to preserve an existing subject, composition, product, or style cue.
Describe the subject, action, environment, framing, lighting, palette, and intended format. Put essential details before decorative ones and avoid conflicting instructions.
Match the canvas to where the image will appear. Choose the supported resolution you need, up to 2K, and request several outputs when comparing creative directions.
Pick the image that best matches your core idea. Ask for one clear revision, review what changed, and continue if needed. Save only after checking details, text, and brand accuracy.
Use cases
Turn a sketch, render, or written description into several polished directions. Compare materials, environments, lighting, and composition before choosing what to develop further.
Create product scenes, social images, hero concepts, icons, and ad variations from a shared visual brief, carrying reference imagery through generation and editing.
Use image-to-image editing and style transfer to test treatments around one composition, from realistic photography to illustration or watercolor.
Build a new visual from a prompt, then use conversational edits to improve the details without learning a long sequence of editing tools first.
Comparison
| Need | Grok Image 2.0 | Standalone Image Editor | Manual Production |
|---|---|---|---|
| Start from a written brief | Generates a new visual from a text prompt | Usually begins with an existing canvas or asset | Requires photography, illustration, or design work |
| Use visual references | Accepts image input, including multi-image editing workflows | Places or edits imported references | Uses references during the production process |
| Revise with plain language | Supports prompt-based image edits and multi-turn refinement | Uses direct editing controls and tool-specific workflows | Changes are recreated by the producer |
| Explore visual styles | Restyles an image from a written instruction | Applies manual adjustments, filters, or generative tools | Requires a new treatment or production pass |
| Finish exact brand details | Requires careful review and may need follow-up work | Gives direct control over fine adjustments | Offers deliberate control but takes production time |
Choose Grok Image 2.0 when speed of exploration and prompt-based revision matter. Use a dedicated editor or manual production step when exact typography, measurements, legal approvals, or pixel-level control are essential.
Why choose
Keep the brief and follow-up edits in the same thread. Point to the current image and describe what should be different next.
Use text when the idea exists only in words, a reference image when the subject already exists, or multiple images when you need separate cues for people, objects, setting, or style.
Set an aspect ratio that fits the placement and select output up to 2K for a square tile, vertical post, or wide page section.
Once you have an approved still, explore a related Grok Imagine video workflow without moving the idea into an unrelated product family.
FAQ
Open Grok Imagine, enter a visual prompt or add a reference photo, and generate an image. Then ask Grok Image 2.0 to edit the strongest result with a focused follow-up instruction.
Yes. Grok Image 2.0 supports text-to-image generation, so you can describe a subject, scene, lighting, composition, and style in writing and receive a generated visual.
Yes. You can provide a source image and describe the change you want in plain language. Multi-turn editing allows one revised image to become the starting point for the next change.
Yes. The supported multi-image editing workflow accepts at least two source images and uses your prompt to describe how their subjects or visual cues should be combined.
Grok Imagine image workflows support output up to 2K. Available choices can depend on the model, interface, and account context.
Yes. Choose the destination first so Grok Image 2.0 can frame the subject for a wide banner, square tile, portrait post, or another supported layout.
Generated images and videos in Grok include a Grok watermark that identifies AI-created content. Grok does not provide a setting to remove it.
Grok is free to try, but usage limits apply. Paid SuperGrok plans raise those limits, and API image generation is billed separately according to the selected model and output.
No. Grok Image 2.0 focuses on image creation and editing, while Grok Imagine also provides separate video generation and editing workflows.
Check people, hands, objects, perspective, spelling, logos, packaging, and factual details. Also confirm that you have the rights and approvals required for every reference image and intended use.
Move from a written idea or reference photo to a visual you can review and refine. Try Grok Imagine Image 2.0, compare several directions, and keep editing until the image matches your brief.
Try Grok Image 2.0