Workspaces
Grok Image 2.0 — AI image generator and editor
One account, one credit balance and a full image studio: generate from text, edit with precision, resize for any channel. The server quotes each task before it starts.
Every task is quoted by the server before it starts, and the ledger keeps the record of what was charged.
Made with Grok Imagine Image 2.0
Real examples from the official xAI Image 2.0 release: photography, posters, infographics, illustration, product edits and consistent world-building.
Source: xAI Imagine Image 2.0 release. Official examples shown for reference.
What is Grok Imagine Image 2.0?
Image 2.0 is the image generation model in the Grok Imagine family from xAI, available in the API as grok-imagine-image-2.0. It is built around one goal: make images you can use in real work. The image workspace on this site runs it in the browser, alongside the other models in the catalogue.
- Instruction following: 2.0 follows the brief closely, down to small details, so dense multi-part visuals hold together.
- Typography and layout: text, posters and infographics come out sharp, with small text that stays readable.
- Targeted editing: edit only the region you point at; segmentation selects precise areas to change while the rest stays untouched.
- Background removal: any subject exports with a transparent background, ready to drop into other work.
- Multi-reference editing: bring up to five reference images together without manual compositing.
- Smart resize: pick a ratio and the model fills the frame — one image for every channel.
What you can create here
The image workspace covers the full loop from a first prompt to a finished asset, and the video workspace extends the same catalogue into motion.
- Text to image: describe the subject, scene, style and placement, then generate.
- Image editing: keep what already works and change one thing at a time — background, color, product or text.
- Product and e-commerce visuals: studio scenes, lifestyle settings and listing photos.
- Posters, infographics and layouts: typography-led images with deliberate structure.
- Video: text to video, image to video and editing with the models listed on this site.
Made for practical creative work
Pick the channel first, generate for that frame, then refine without restarting.
- Marketing and campaigns: launch creative, banners and advertising concepts around one clear message.
- Social content: square posts, vertical stories, thumbnails and wide banners, sized as you generate.
- Brand and identity assets: portraits, icons, mascots and emoji that stay consistent across versions.
- Illustration and concept art: characters, environments and mood across photographic and illustrated styles.
How a task is billed
Every workspace draws on the same credit balance. The server quotes the cost of a task before it starts, and the ledger keeps the record of what was actually charged.
- Pick the workspace and the model.
- Read the quoted credit cost for the settings you chose.
- Submit the task and follow it to completion.
- Download the finished file from the workspace.
Explore by use case
Posters & infographics
Travel posters, explainers and study sheets that read at a glance.
Cinematic photography
Portraits and scenes with consistent lighting, grading and mood.
Product & brand visuals
Product shots, template covers and campaign-ready marketing imagery.
Characters & world building
Consistent characters, props and environments across a whole series.
Examples

Cinematic fantasy photography

Text-heavy infographic with readable small type

Poster design with deliberate layout

Photorealistic close-up

Pixel art and game assets

Consistent world-building across generations

Product color change without reshooting

Portrait refined into a clean headshot
Questions
What is Grok Imagine Image 2.0?
Image 2.0 is xAI's image generation and editing model in the Grok Imagine family, available in the API as grok-imagine-image-2.0. This site runs it in the image workspace together with the other models in the catalogue.
Can I edit an existing image instead of starting over?
Yes. Image 2.0 is built for targeted edits: point at the region that needs to change, keep the rest of the composition, replace or remove backgrounds, or restyle a scene without losing the subject.
Do I need an account to generate?
Yes. Signing in is required before a task can be submitted, and your credits, orders and finished work are kept in that account.
Where does the credit price come from?
The server quotes each task from the active pricing rules before it starts. The quotation and the ledger are the billing record; prices written on a page are list prices only.
What kinds of images can I create?
Photography-style visuals, product concepts, marketing graphics, posters, infographics, social content, illustration, icons and concept art. Choose the aspect ratio for the channel where the image will be used.
Which languages does the site offer?
Six: English, Simplified Chinese, Traditional Chinese, German, Spanish and Portuguese. Each page exists in the languages listed in its language menu.
Explore next
AI image generator
Describe an image or upload a reference, choose a model, and read the quoted credit cost before generating.
AI video generator
Write a shot, animate a still, or drive a clip with a reference video, then read the quoted credit cost before the job starts.
Image to prompt
Upload a picture, get a written description you can reuse as a prompt. Five target models, three detail levels, optional negative prompt, four output languages.
Grok Imagine Image 2.0
Grok Imagine Image 2.0 as this workspace declares it: sixteen aspect ratios, 1k to 4k, up to fourteen reference images. All cases are third-party video frames.
Nano Banana 2.1
Nano Banana 2.1 is Google's newest image model, GA since 6 October 2026. What the publisher documents, what this site's catalogue declares for the product, and how it compares with Nano Banana 2 and Nano Banana 2 Lite on this workspace.
GPT Image 2.5 Flare
GPT Image 2.5 Flare as this catalogue declares it: the modes it accepts, the canvas it allows, what the four-point tier buys, and what the published artwork cannot prove.
GPT Image 2.5 Sunburst
GPT Image 2.5 Sunburst in the publisher's own words, the single official figure that separates it from Flare, and why this workspace declares identical controls for both.
Nano Banana 2
Nano Banana 2 is Gemini 3.1 Flash Image. What Google published about it, what this catalogue declares, and the controls the publisher ships that this workspace form does not carry.
Nano Banana 2 Lite
Nano Banana 2 Lite is Gemini 3.1 Flash-Lite Image. What the publisher published about it, what this catalogue returns for it today, and why the workspace form cannot send a task to it yet.
Nano Banana Pro
Generate and edit images with Nano Banana Pro, Google's Gemini 3 Pro Image model. See what this workspace declares, 18 published samples, and start a job.
Doubao Seedream 5.0 Pro
Seedream 5.0 Pro as this catalogue declares it: three task types, ten reference images and a free-text size field, against an API that documents thirteen parameters.
FLUX 3 Image
FLUX 3 Image on this site: the Black Forest Labs layout model with bounding boxes, ten reference images and native 4K, as the site-19 catalogue actually exposes it.
Grok Imagine Video 1.5
Grok Imagine Video 1.5 as the publisher documents it set beside what this site's service actually publishes: modes, duration tiers, resolutions and reference limits.
Gemini Omni 1.1 Flash
Gemini Omni 1.1 Flash as Google documents it beside what this site's service publishes: five declared modes, four the interface can express, and per-second prices for four tiers.
Seedance 2.5
Seedance 2.5 is ByteDance's longer-form video model. Here are the 4 to 30 second lengths, 480p to 1080p qualities and reference modes this workspace declares, plus 16 published samples.
Seedance 2.0
Seedance 2.0 is ByteDance's video model. Here are the modes, 2-15 second lengths, 480p to 1080p qualities and audio toggle this workspace declares, plus 17 published samples.
Seedance 2.0 Fast
Seedance 2.0 Fast is Seedance 2.0 without the 1080p setting. This page sets out the declared fields, compares all three tiers, and explains why every sample is labelled 2.0.
Vidu Q3
Vidu Q3 as Shengshu Technology documents it, set beside the single mode this site's service publishes for it, with the publisher's own clips and the request-example inputs kept apart.
Wan 3.0
Wan 3.0 as Alibaba Cloud documents it, beside the three modes this site's service publishes for it, with the reference-asset limits, the resolution default and the width the form cannot reach all stated plainly.
HappyHorse 1.1
HappyHorse 1.1 (快乐小马) has three published modes and nine reference-image slots. This page records which of them this site's own workspace can actually reach, read from the source, and where every picture on it came from.
Guides
Step-by-step guides for the tasks this site supports, written around the workspaces that are actually mounted here.