> ## Documentation Index
> Fetch the complete documentation index at: https://docs.qingbo.dev/llms.txt
> Use this file to discover all available pages before exploring further.

# GPT Image Series

> Four GPT Image models from OpenAI: three official lines billed on token usage, one reverse line at a fixed price per image

All four models in the GPT Image series are served on `POST /v1/tasks` — the request URL does not change between them. They share the same actions and request shape; the differences are the aspect-ratio and resolution values, the line-specific parameters, and billing. Picking a model only changes the `model` field.

## Available models

<CardGroup cols={2}>
  <Card title="GPT Image 1" href="/en/api-reference/image/gpt-image/gpt-image-1">
    `gpt-image-1` · text-to-image, edits with up to 15 references, mask inpainting and transparency; billed on actual token usage
  </Card>

  <Card title="GPT Image 1.5" href="/en/api-reference/image/gpt-image/gpt-image-1">
    `gpt-image-1.5` · the same actions and parameters as `gpt-image-1`, with text output as an extra billing dimension; billed on actual token usage
  </Card>

  <Card title="GPT Image 2 Official" href="/en/api-reference/image/gpt-image/gpt-image-2">
    `gpt-image-2` · 15 aspect ratios, `1k` / `2k` / `4k` resolutions, up to 16 references, mask inpainting and WebP output; billed on actual token usage
  </Card>

  <Card title="GPT Image 2 Reverse" href="/en/api-reference/image/gpt-image/gpt-image-2-rev">
    `gpt-image-2-rev` · text-to-image and edits with up to 15 references; fixed per-image pricing at `1k`, `2k`, or `4k`
  </Card>
</CardGroup>

## Choose a model

| Need                                                               | Recommended model                           | Why                                                                                  |
| ------------------------------------------------------------------ | ------------------------------------------- | ------------------------------------------------------------------------------------ |
| `2k` / `4k` output, or an aspect ratio beyond `1:1` `2:3` `3:2`    | `gpt-image-2`                               | It is the only model that accepts `resolution`, and it supports 15 aspect ratios     |
| A fixed per-image price before submission                          | `gpt-image-2-rev`                           | Billing is fixed by output resolution, independent of aspect ratio and prompt length |
| Mask inpainting, transparent backgrounds, or output-format control | `gpt-image-1` `gpt-image-1.5` `gpt-image-2` | All three official lines expose these line-specific parameters                       |
| WebP output                                                        | `gpt-image-2`                               | The other three models output `png` and `jpeg` only                                  |
| Multi-image reference editing                                      | All four                                    | `gpt-image-2` accepts up to 16 images; the other three accept up to 15               |

<Note>
  The models never switch automatically: the `model` field selects the model, and its own page is authoritative for the parameters it accepts.
</Note>

## What all four share

These are identical on all four models:

| Item                 | Value                                                                             |
| -------------------- | --------------------------------------------------------------------------------- |
| Endpoint             | `POST /v1/tasks` (async task, see [Task system](/en/docs/task-system))            |
| Actions              | `generate` text-to-image · `edit` reference-image editing (requires `image_urls`) |
| Default aspect ratio | `1:1`; the available set differs per model, see each model page                   |
| Output count         | One image per request                                                             |
| Not supported        | `seed`                                                                            |

## Unified request

```bash theme={"system"}
curl -X POST https://www.qingbo.dev/v1/tasks \
  -H "Authorization: Bearer $WAVE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-image-2",
    "action": "generate",
    "prompt": "A blue ceramic cup on a white table in soft natural light",
    "aspect_ratio": "1:1",
    "n": 1
  }'
```

Swap `model` for any model ID on this page and the request still runs. A successful submission returns a `task_id`. Retrieve the result with [`GET /v1/tasks/{task_id}`](/en/api-reference/task/status), wait inline with [`Prefer: wait`](/en/api-reference/task/submit#in-request-waiting-prefer-wait), or configure a webhook.

## Related

* [GPT Image 1 and 1.5 Official](/en/api-reference/image/gpt-image/gpt-image-1)
* [GPT Image 2 Official](/en/api-reference/image/gpt-image/gpt-image-2)
* [GPT Image 2 Reverse](/en/api-reference/image/gpt-image/gpt-image-2-rev)
* [Submit Task](/en/api-reference/task/submit)
* [Query Task Status](/en/api-reference/task/status)
* [Task System](/en/docs/task-system)