Image ● Operational Service status →

Grok Imagine Image

⧉

grok-imagine-image

Model Type Standard Quality

Grok Imagine Image is xAI's image generation model (Standard tier). One model handles text-to-image and image editing: 1K usually renders in 5-10 seconds and 2K in about 12, with accurate Chinese and English lettering. Edits can take several reference images and repaint while keeping the original subject. OpenAI images API compatible: set model to grok-imagine-image. 1K and 2K cost the same; billed per image, failed jobs refunded in full.

Text-to-imageImage editing1K/2KFast

Get API Key → API Docs

Live Test · Playground

Try out Grok Imagine Image right here (available after login).

Model Type

Input

Image

Prompt

Aspect ratio

Quality

Output

image

Results will appear here after running.

🕑 Results are kept for 60 days and then deleted automatically, so download anything you want to keep.

About Grok Imagine Image

Grok Imagine Image is xAI's image generation model; NezhaGate offers the Standard tier (grok-imagine-image) and the Quality tier (grok-imagine-image-quality). Standard is about speed and unit price: 1K usually renders in 5-10 seconds and 2K in about 12, with accurate Chinese and English lettering, which suits posters, social visuals and batch work. The API is compatible with the OpenAI images API: text-to-image goes to /v1/images/generations, and for an edit you put the references in image or images (a URL, a data: URI or base64) or call /v1/images/edits. Submitting returns a job id at once; poll GET /v1/images/jobs/{id} for the image. Seven aspect ratios (1:1, 3:4, 2:3, 9:16, 4:3, 3:2, 16:9) and two resolutions, 1K and 2K; 4K is not offered. 1K and 2K cost the same, billed per image, with failed jobs refunded in full.

Use cases

Posters and campaign art

Headlines, slogans and prices in Chinese or English go straight into the prompt, and renders come back fast, for quick iterations on posters and product banners.

Social and content visuals

Covers and images for Xiaohongshu, WeChat, X and the like, rendered directly at 3:4, 9:16, 16:9 and other ratios.

Image editing

New backgrounds, new materials, a logo turned into a physical object: the original subject stays and the rest is repainted to the prompt; several references can be combined.

Batch generation

A low unit price and fast renders suit workflows that generate many candidates and pick the best.

How to choose

Choose Standard, grok-imagine-image, for speed, unit price and batch work; choose Quality, grok-imagine-image-quality, for finished covers and posters where detail and lettering matter more. Quality costs twice as much as Standard, again with 1K and 2K at the same price. For 4K, use GPT Image 2 or the Nano Banana models.

FAQ

Which aspect ratios and resolutions are supported?

Seven aspect ratios - 1:1, 3:4, 2:3, 9:16, 4:3, 3:2, 16:9 - at 1K or 2K (send resolution: "2K" or a 2048-class pixel size); 4K is not offered. The gateway steers the model to compose at the chosen ratio and then crops to the exact size you asked for.

Does an edit change the aspect ratio?

On the Standard tier an edit keeps the aspect ratio of the original image; if you ask for a different ratio the gateway centre-crops to it, so content at the edges can be cut. For edits, set size to the same ratio as the original, or use the Quality tier, which renders at the ratio you choose.

How many reference images can I send?

Several: one in image, or several in images (in our tests two references were combined into a single picture). Each can be a public URL, a data: URI or base64.

Am I charged when a render fails?

No. Failed renders, timeouts and content-moderation refusals are refunded in full; you pay per image only when one is delivered.

Related Models

Explore other models you can integrate.

View all →

Image