Grok Imagine Image 2.0 API: Image Generation & Editing

Overview

Grok Imagine Image 2.0 supports text-to-image generation and image-to-image editing. Each successful task returns one image.

Capability	Value
Model ID	`grok-imagine-image-2-0`
Modes	`text-to-image`, `image-to-image`
Prompt	Required for text-to-image; optional for image-to-image; up to 8,000 characters
Reference images	1–5 URLs for image-to-image
Aspect ratios	`1:1`, `2:3`, `3:2`, `16:9`, `9:16`; image-to-image also supports `auto`
Output	One image URL in `resultJson.resultUrls`

Endpoint and authentication

Base URL:

https://api.apixo.ai/api/v1

Method	Endpoint	Purpose
`POST`	`/generateTask/grok-imagine-image-2-0`	Submit a generation or editing task
`GET`	`/statusTask/grok-imagine-image-2-0?taskId={taskId}`	Poll task status and retrieve the result

Include Authorization: Bearer YOUR_API_KEY on every request and Content-Type: application/json when submitting JSON.

Copy-paste async quickstart

curl -X POST "https://api.apixo.ai/api/v1/generateTask/grok-imagine-image-2-0" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "request_type": "async",
    "input": {
      "mode": "text-to-image",
      "prompt": "A small wooden cabin beside a lake at sunrise",
      "aspect_ratio": "16:9"
    }
  }'

An accepted task returns a task ID. Acceptance does not mean generation is complete.

{
  "code": 200,
  "message": "success",
  "data": {"taskId": "task_12345678"}
}

Image-to-image

Provide 1–5 publicly accessible reference URLs. The prompt is optional; include it to describe the desired edit. Omitting aspect_ratio uses auto.

{
  "request_type": "async",
  "input": {
    "mode": "image-to-image",
    "image_urls": ["https://example.com/reference.png"],
    "prompt": "Keep the subject and change the background to a snowy mountain",
    "aspect_ratio": "auto"
  }
}

Parameters

string

default:"async"

Result delivery mode. Supported values: async, callback. Omit this field or use async to poll for the result.

string

Required when request_type is callback. Provide a public HTTPS URL that receives the final task payload. See Webhooks.

object

required

Grok Imagine Image 2.0 input parameters.

Show properties

string

required

Generation mode. Supported values: text-to-image, image-to-image.

string

Required and non-blank for text-to-image; optional for image-to-image. Maximum 8,000 characters. For image editing, omit it or provide an edit instruction.

string[]

Required for image-to-image: 1–5 publicly accessible image URLs, each a non-empty string. Preserve the intended reference order. Ignored for text-to-image.

string

Output aspect ratio. Supported values: 1:1, 2:3, 3:2, 16:9, 9:16, auto. The value auto is supported only for image-to-image. Defaults to 1:1 for text-to-image and auto for image-to-image.

Poll for result

curl "https://api.apixo.ai/api/v1/statusTask/grok-imagine-image-2-0?taskId=task_12345678" \
  -H "Authorization: Bearer YOUR_API_KEY"

Continue polling while data.state is pending or processing. Stop at success or failed. Successful response:

{
  "code": 200,
  "message": "success",
  "data": {
    "taskId": "task_12345678",
    "state": "success",
    "resultJson": "{\"resultUrls\":[\"https://example.com/generated.png\"]}",
    "createTime": 1793318400000,
    "completeTime": 1793318440000,
    "costTime": 40000
  }
}

resultJson is a JSON string, so parse it before reading the image URL:

const { resultUrls } = JSON.parse(response.data.resultJson);
const imageUrl = resultUrls[0];

Status response fields

Field in `data`	Type	Description
`taskId`	string	Task identifier returned at submission
`state`	string	`pending`, `processing`, `success`, or `failed`
`resultJson`	string	JSON string containing `resultUrls` on success
`failCode`	string	Failure code when the task fails
`failMsg`	string	Failure message when the task fails
`createTime`	integer	Task creation timestamp in Unix milliseconds
`completeTime`	integer	Task completion timestamp in Unix milliseconds, when available
`costTime`	integer	Total task duration in milliseconds, when available

Failed task example

{
  "code": 200,
  "message": "success",
  "data": {
    "taskId": "task_12345678",
    "state": "failed",
    "failCode": "500",
    "failMsg": "Image generation failed."
  }
}

A successful status lookup can describe a failed generation. Check data.state, and retain the task ID when requesting support.

Webhook callback mode

Set request_type to callback and provide callback_url; the model input is unchanged.

{
  "request_type": "callback",
  "callback_url": "https://your-server.com/webhooks/apixo",
  "input": {
    "mode": "text-to-image",
    "prompt": "An illustrated travel poster of a quiet seaside town",
    "aspect_ratio": "2:3"
  }
}

Webhook delivery uses the same public final task fields as polling. See Webhooks for delivery and retry requirements.

Billing

Both modes cost $0.04 per generated image, with one output image per successful task. Reference images do not add an input charge. Failed tasks are refunded.

Mode	APIXO price	Market display reference
`text-to-image`	`$0.040 / image`	`$0.060 / image`
`image-to-image`	`$0.040 / image`	`$0.060 / image`

The market display reference is a comparison value and does not affect the amount charged. New non-LLM requests use 1 credit = $0.001. Calculate the full request amount, apply any applicable account discount, then round the total upward to the next whole credit. Agreed fixed unit prices retain their precision and replace the standard price and discount; final rounding still applies. Without discounts, one successful task costs 40 credits. See credit conversion and rounding and current API prices.

Latency and polling

Generation time varies with prompt complexity, image access, and queue load. The following intervals are client recommendations, not latency guarantees.

Mode	Recommended first poll	Poll interval
`text-to-image`	10 seconds after task creation	3–5 seconds, with backoff
`image-to-image`	10 seconds after task creation	3–5 seconds, with backoff

Use callback mode for completion notifications. Stop polling at a terminal state, and save important outputs promptly.

Errors and troubleshooting

Missing mode, a missing or blank text-to-image prompt, an overlong prompt, an invalid aspect ratio, or an invalid image-to-image reference list causes a validation error.
Use reference URLs that are accessible without authentication. If generation fails, inspect failCode and failMsg.
Check your API key for authentication errors and your balance for insufficient-balance errors.
Back off on rate limits. Do not resubmit an accepted task just because it is still processing; continue polling its task ID.

See Error Codes for shared API errors.

MCP integration

The complete model schema declares result decoding in mcp.result and supported file bindings in mcp.attachments. These are client integration metadata, not generation input parameters. Attachment path: image_urls. Preserve the user’s file order. Model parameter validation remains on the API.

​Overview

​Endpoint and authentication

​Copy-paste async quickstart

​Image-to-image

​Parameters

​Poll for result

​Status response fields

​Failed task example

​Webhook callback mode

​Billing

​Latency and polling

​Errors and troubleshooting

​MCP integration

​Related links

Overview

Endpoint and authentication

Copy-paste async quickstart

Image-to-image

Parameters

Poll for result

Status response fields

Failed task example

Webhook callback mode

Billing

Latency and polling

Errors and troubleshooting

MCP integration

Related links