GPT Image 2.5 Sunburst

POST Base URL: https://api.hiapi.ai /v1/tasks

Image, video, and audio models are called through the Unified Async API POST /v1/tasks endpoint; only the input fields differ (see input parameters below).

Model summary

Model name	gpt-image-2.5-sunburst
Type	Text-to-image / Image editing
Output	One image per request
Pricing	View current pricing

GPT Image 2.5 Sunburst supports text-to-image and image editing through one model ID, with up to 16 reference images, six quality settings, and aspect-ratio or pixel-size selection.

Production guidance

Image inputs and outputs

Omit image_urls for prompt-only generation. Supply image URLs for editing or references, and do not send an empty array.
Each request generates one image. Download or copy the result to your storage before its temporary URL expires.
Use PNG or WebP for transparency and describe the subject and background requirements in your prompt.

Best suited for

Product and scene visuals

Describe the subject, material, composition, and lighting to generate a complete image from text.

promptaspect_ratio

Editing and multiple references

Supply subjects or style through reference images and specify what to change and preserve.

promptimage_urls

Request parameters

POST https://api.hiapi.ai/v1/tasks

# Create an async task — returns taskId immediately (generation runs async)
# Replace YOUR_API_KEY with your account API key
curl -X POST "https://api.hiapi.ai/v1/tasks" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "gpt-image-2.5-sunburst",
  "input": {
    "prompt": "Change only the central crescent-moon inlay on the open lacquer cabinet into a perfectly circular total lunar eclipse disc made from deep crimson mother-of-pearl with a thin copper-gold halo. Preserve every surrounding shell fragment, gold-wire constellation, cabinet shape, door angle, brass hinge, conservation tool, tray, lacquer reflection, camera position, focus, lighting direction, shadow, and background exactly. Do not add, remove, move, resize, or recolor anything else. No text, logos, people, animals, perfume, or watches.",
    "image_urls": [
      "https://example.com/reference.webp"
    ],
    "aspect_ratio": "1:1",
    "quality": "high",
    "background": "opaque",
    "output_format": "webp"
  }
}'

import requests, json

# Business params — fields vary by model (see the model page)
payload = json.loads(r'''
{
  "model": "gpt-image-2.5-sunburst",
  "input": {
    "prompt": "Change only the central crescent-moon inlay on the open lacquer cabinet into a perfectly circular total lunar eclipse disc made from deep crimson mother-of-pearl with a thin copper-gold halo. Preserve every surrounding shell fragment, gold-wire constellation, cabinet shape, door angle, brass hinge, conservation tool, tray, lacquer reflection, camera position, focus, lighting direction, shadow, and background exactly. Do not add, remove, move, resize, or recolor anything else. No text, logos, people, animals, perfume, or watches.",
    "image_urls": [
      "https://example.com/reference.webp"
    ],
    "aspect_ratio": "1:1",
    "quality": "high",
    "background": "opaque",
    "output_format": "webp"
  }
}
''')

# Create an async task — returns taskId immediately (generation runs async)
resp = requests.post(
    "https://api.hiapi.ai/v1/tasks",
    headers={"Authorization": "Bearer YOUR_API_KEY", "Content-Type": "application/json"},
    json=payload,
)
print(resp.json())  # Returns { code, message, data: { taskId } }

// Create an async task — returns taskId immediately (generation runs async)
// Replace YOUR_API_KEY with your account API key
const resp = await fetch("https://api.hiapi.ai/v1/tasks", {
  method: "POST",
  headers: {
    "Authorization": "Bearer YOUR_API_KEY",
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
  "model": "gpt-image-2.5-sunburst",
  "input": {
    "prompt": "Change only the central crescent-moon inlay on the open lacquer cabinet into a perfectly circular total lunar eclipse disc made from deep crimson mother-of-pearl with a thin copper-gold halo. Preserve every surrounding shell fragment, gold-wire constellation, cabinet shape, door angle, brass hinge, conservation tool, tray, lacquer reflection, camera position, focus, lighting direction, shadow, and background exactly. Do not add, remove, move, resize, or recolor anything else. No text, logos, people, animals, perfume, or watches.",
    "image_urls": [
      "https://example.com/reference.webp"
    ],
    "aspect_ratio": "1:1",
    "quality": "high",
    "background": "opaque",
    "output_format": "webp"
  }
}),
});
console.log(await resp.json()); // Returns { code, message, data: { taskId } }

model string required

Select this model.

example gpt-image-2.5-sunburst

input object required

Image generation or editing parameters.

prompt string required

Describe the image to generate; for edits, specify what to change and preserve. Use 1–32000 characters.

image_urls string[] optional

Image URLs for editing or visual reference. When provided, include 1–16 images. Omit this field for text-to-image; for multiple images, explain each image’s role in array order.

aspect_ratio enum optional

Choose an output aspect ratio or an explicit width-by-height pixel size. auto lets the model choose a suitable composition; pixel sizes use a lowercase x separator.

default 1:1

Aspect ratio presets

1:13:22:34:33:416:99:16auto

Pixel dimensions

1024x10241536x10241024x15361536x11521152x15362048x20482048x11521152x20483840x21602160x3840

quality enum optional

Choose image quality. low, medium, high, xhigh, and max provide different detail levels; auto lets the model decide.

default medium enum: lowmediumhighxhighmaxauto

background enum optional

Background treatment: auto chooses automatically, transparent requests transparency, and opaque requests a nontransparent background.

default auto enum: autotransparentopaque

output_format enum optional

File format for the result. Choose PNG or WebP for transparent backgrounds; JPEG does not support transparency.

default webp enum: pngjpegwebp

callback object optional

Configure a notification when the task finishes.

url string required

HTTPS URL that receives terminal task notifications.

example https://your-domain.com/hiapi/callback

when enum optional

Notify when the task succeeds or fails.

default final enum: final

Example requests

Precise Lunar Lacquer Edit

Replace the example URL with your lacquer cabinet image and change only the central lunar inlay while preserving the surrounding decoration, cabinet, and lighting.

Request body

{
  "model": "gpt-image-2.5-sunburst",
  "input": {
    "prompt": "Change only the central crescent-moon inlay on the open lacquer cabinet into a perfectly circular total lunar eclipse disc made from deep crimson mother-of-pearl with a thin copper-gold halo. Preserve every surrounding shell fragment, gold-wire constellation, cabinet shape, door angle, brass hinge, conservation tool, tray, lacquer reflection, camera position, focus, lighting direction, shadow, and background exactly. Do not add, remove, move, resize, or recolor anything else. No text, logos, people, animals, perfume, or watches.",
    "image_urls": [
      "https://example.com/reference.webp"
    ],
    "aspect_ratio": "1:1",
    "quality": "high",
    "background": "opaque",
    "output_format": "webp"
  }
}

Lunar Lacquer Cabinet Source

Generate a lunar lacquer cabinet in a conservation studio, with mother-of-pearl inlay, gold-wire constellations, and black lacquer detail.

Request body

{
  "model": "gpt-image-2.5-sunburst",
  "input": {
    "prompt": "Create a museum-grade square still-life photograph inside a quiet lacquer conservation studio. Center an open, deep black urushi lacquer jewelry cabinet on a charcoal conservation table. Its doors and interior panels contain an intricate lunar map made from hundreds of tiny iridescent mother-of-pearl fragments, hair-thin gold wire constellations, miniature brass hinges, and hand-polished black lacquer with controlled mirror reflections. Around it place only precise archival tools: fine tweezers, a sable brush, a linen glove, and two labeled-free shell trays. A single warm raking task lamp reveals nacre rainbow colors and minute surface texture against an obsidian background. Symmetrical yet natural composition, exceptional micro-detail, restrained black, pearl, teal, violet, and champagne-gold palette, luxury craft editorial photography, no people, no animals, no perfume, no watches, no text, no logos, no watermark.",
    "aspect_ratio": "1:1",
    "quality": "high",
    "background": "opaque",
    "output_format": "webp"
  }
}

Getting the result

Submission returns a taskId while the image is generated asynchronously.
Poll GET /v1/tasks/:id or wait for the terminal notification at callback.url.
When status=success, save the image from data.output[].url before expireAt.
When status=fail, inspect data.error and correct the request before deciding whether to submit again.

FAQ

Do generation and editing use different model IDs?

No. Use the model ID on this page. Omit image_urls for text-to-image, or provide images for editing or visual reference.

How should I organize multiple references?

Describe each image’s role in image_urls order, such as the first for the product and the second for the scene, and specify which elements should remain unchanged.

How do I choose dimensions?

Choose an aspect ratio for common layouts, or select an explicit pixel size from the width-by-height values listed under aspect_ratio. Other custom dimensions are not accepted.

Do I need a separate API key for each model?

Use an API key created in your HiAPI dashboard and send it as a Bearer token. Keep the key on your server.