Contratto di integrazione GPT Image 2.6 | Documentazione API

# GPT Image 2.6

GPT Image 2.6 Flare and Sunburst integration contract.

POST https://api.apihop.ai/openai/v1/images/generations

Integration is prepared for `gpt-image-2.6-flare` and `gpt-image-2.6-sunburst`. APIMart documentation still labels this upstream API an integration preview. Availability depends on your current API key and gateway activation; check `GET https://api.apihop.ai/v1/models`.

## Request

```http
POST https://api.apihop.ai/openai/v1/images/generations
Authorization: Bearer YOUR_API_KEY
Content-Type: application/json
```

```json
{
  "model": "gpt-image-2.6-flare",
  "prompt": "A red ceramic cup on a white table.",
  "size": "1:1",
  "resolution": "1k",
  "quality": "low",
  "output_format": "png",
  "n": 1
}
```

## Parameters

| Parameter | Values |
| --- | --- |
| `model` | `gpt-image-2.6-flare`, `gpt-image-2.6-sunburst` |
| `prompt` | Required text |
| `n` | 1–4 |
| `size` | `auto`, aspect ratio, `WIDTHxHEIGHT` |
| `resolution` | `1k`, `2k`, `4k` |
| `quality` | `auto`, `low`, `medium`, `high`, `xhigh`, `max` |
| `image_urls` | 0–16 HTTP(S) URLs |
| `output_format` | `png`, `jpeg`, `webp` |
| `output_compression` | 0–100; JPEG/WebP |
| `background` | `auto`, `opaque`, `transparent`; PNG/WebP |
| `complexity` | `low`, `medium`, `high` |
| `safety_identifier` | String, maximum 128 characters |
| `stream` | `false` (JSON), `true` (SSE) |

Pixel dimensions must be multiples of 16, with each side at most 3840, an aspect ratio at most 3:1, and 655,360–8,294,400 total pixels. Reference editing uses `image_urls` with public URLs on the same endpoint. This initial integration excludes base64 references, multipart uploads, masks, markup, GIF, SVG, animation, and `/images/edits`.

## Response

The connection waits for completion. JSON returns final URLs in `data[].url` and available token `usage`. SSE completes with `image_generation.completed` and `[DONE]`; partial image previews are not guaranteed. Clients do not poll the upstream task. Do not resubmit requests whose completion status is unknown.

## Pricing

Billing separates text input, cached text input, image input, cached image input, and image output tokens. Check the current catalog price; temporary reservations are reconciled after completion. Published APIMart reference prices do not prove account costs or live availability.

L’integrazione è pronta per gpt-image-2.6-flare / gpt-image-2.6-sunburst. La documentazione APIMart descrive ancora questa API come anteprima di integrazione. La disponibilità dipende dalla chiave API attuale e dall’attivazione del gateway; controlla GET https://api.apihop.ai/v1/models.

Richiesta

http

POST https://api.apihop.ai/openai/v1/images/generations
Authorization: Bearer YOUR_API_KEY
Content-Type: application/json

json

{
  "model": "gpt-image-2.6-flare",
  "prompt": "A red ceramic cup on a white table.",
  "size": "1:1",
  "resolution": "1k",
  "quality": "low",
  "output_format": "png",
  "n": 1
}

Parametri

Parametro	Valori
`model`	`gpt-image-2.6-flare`, `gpt-image-2.6-sunburst`
`prompt`	Testo obbligatorio
`n`	1–4
`size`	`auto`, rapporto d’aspetto, `WIDTHxHEIGHT`
`resolution`	`1k`, `2k`, `4k`
`quality`	`auto`, `low`, `medium`, `high`, `xhigh`, `max`
`image_urls`	0–16 HTTP(S) URLs
`output_format`	`png`, `jpeg`, `webp`
`output_compression`	0–100; JPEG/WebP
`background`	`auto`, `opaque`, `transparent`; PNG/WebP
`complexity`	`low`, `medium`, `high`
`safety_identifier`	Stringa, massimo 128 caratteri
`stream`	`false` (JSON), `true` (SSE)

Le dimensioni devono essere multipli di 16, ogni lato al massimo 3840, rapporto massimo 3:1 e totale 655,360–8,294,400 pixel. I riferimenti usano image_urls con URL pubblici sullo stesso endpoint. Questa integrazione esclude riferimenti base64, upload multipart, maschere, markup, GIF, SVG, animazioni e /images/edits.

Risposta

La connessione attende il completamento. JSON restituisce URL finali in data[].url e token usage disponibili. SSE termina con image_generation.completed e [DONE]; anteprime parziali non garantite. I client non interrogano il task upstream. Non reinviare richieste con esito sconosciuto.

Prezzi

La fatturazione separa token di testo in ingresso, testo in cache, immagine in ingresso, immagine in cache e immagine in uscita. Consulta il prezzo corrente del catalogo; le riserve vengono riconciliate al termine. I prezzi di riferimento APIMart non dimostrano costi dell’account o disponibilità.