# gpt-image-1 in 2026: what replaced it, sizes, prices and prompts

> OpenAI shut down gpt-image-1 on 23 Oct 2026. What GPT Image 2 and 2.5 do (editing, up to 3840 px, transparency), per-image API prices, free use in ChatGPT, 7 prompts.

URL: https://superduperai.co/blog/gpt-image-1/
Updated: 2026-11-02
Published: 2025-05-20

**gpt-image-1 no longer works.** OpenAI shut the model down in its API on 23 October 2026, after announcing it in April. The replacements are **GPT Image 2** (April 2026) and **GPT Image 2.5** in two variants, Sunburst for quality and Flare for speed (September 2026). In ChatGPT the same family runs as "ChatGPT Images 2.5", including on the free plan.

Two things the 2025 version of this page got wrong, and which people still search for: gpt-image-1 **could** edit uploaded images (with masks and reference images, from day one), and it **never** made 4096×4096 images. Its only sizes were 1024×1024, 1024×1536 and 1536×1024.

## From gpt-image-1 to GPT Image 2.5

| Date | Model | What changed |
| --- | --- | --- |
| 23 Apr 2025 | **gpt-image-1** | Generation and editing in the API |
| 6 Oct 2025 | gpt-image-1-mini | Cheaper version |
| 16 Dec 2025 | gpt-image-1.5 | Better quality, same sizes |
| 21 Apr 2026 | **gpt-image-2** | Custom sizes, high-fidelity image inputs |
| 8 Sep 2026 | **gpt-image-2.5-sunburst / -flare** | New `xhigh` and `max` quality; better precise edits |
| 23 Oct 2026 | gpt-image-1 shut down | |
| 1 Dec 2026 | gpt-image-1-mini, gpt-image-1.5, chatgpt-image-latest shut down | |

**Migrating:** change `gpt-image-1` to `gpt-image-2` for the closest behaviour, or to `gpt-image-2.5-flare` if speed matters. On GPT Image 2, `input_fidelity` is always high, so edit requests use more input tokens than before.

## What the current models can do

- **Generate and edit.** Send several input images and say what each one is for (subject, style, clothing, background). A mask with an alpha channel marks the area to change; OpenAI notes the edit follows the mask's shape approximately, not pixel for pixel.
- **Sizes.** 1024×1024, 1536×1024 and 1024×1536 are recommended, but any `WIDTHxHEIGHT` works if both sides are multiples of 16, the long side is at most 3840 px and the ratio at most 3:1. So 2048×2048, 2048×1152 and 3840×2160 are fine; anything above 2560×1440 is labelled experimental. 4096×4096 is not valid for any GPT Image model.
- **Quality:** low, medium, high and auto; GPT Image 2.5 adds `xhigh` and `max`.
- **Transparent background:** `background: "transparent"` with PNG or WebP output (gpt-image-1 had it too; GPT Image 2 got it in preview in August 2026).
- **Formats:** PNG, JPEG, WebP, with compression for JPEG and WebP. JPEG is the fastest.
- **Multi-turn editing** through the Responses API, and streaming of up to 3 partial images.

OpenAI's own list of weak spots: complex prompts can take up to two minutes, text placement can still be off, and a character or brand can drift between generations.

## Using it for free in ChatGPT

ChatGPT Images 2.5 is available on every ChatGPT plan, Free included; paid plans get more generations. OpenAI does not publish the free daily limit. Type what you want, or attach a photo and say what to change and what to keep. The API has no free tier for images.

## API prices

Per image, from OpenAI's image generation guide (1024×1024 / 1024×1536 / 1536×1024):

| Model | Low | Medium | High |
| --- | --- | --- | --- |
| GPT Image 2 | $0.006 / $0.005 / $0.005 | $0.053 / $0.041 / $0.041 | $0.211 / $0.165 / $0.165 |
| GPT Image 1.5 (until 1 Dec 2026) | $0.009 / $0.013 / $0.013 | $0.034 / $0.05 / $0.05 | $0.133 / $0.20 / $0.20 |
| GPT Image 1 Mini (until 1 Dec 2026) | $0.005 / $0.006 / $0.006 | $0.011 / $0.015 / $0.015 | $0.036 / $0.052 / $0.052 |
| GPT Image 1 (retired) | $0.011 / $0.016 / $0.016 | $0.042 / $0.063 / $0.063 | $0.167 / $0.25 / $0.25 |

GPT Image 2.5 Sunburst and Flare share one token price: $5 per million text input tokens, $8 per million image input tokens, **$30 per million image output tokens**, half that through the Batch API. OpenAI publishes no per-image figure for 2.5; Artificial Analysis estimates about $0.21 per image at `max`.

## How to prompt GPT Image

Condensed from OpenAI's GPT Image 2.5 prompting guide:

1. **Say what the image is for:** product photo, ad, diagram, thumbnail. Then composition and aspect ratio.
2. **Any format works:** a sentence, a paragraph, labelled sections (scene, subject, details, constraints) or JSON-like text. Pick what you can maintain.
3. **Describe what is visible:** materials, light, colour, medium. Write "photorealistic" if you want a photo.
4. **People:** framing, gaze, how hands hold things ("full body visible, feet included").
5. **Exact text in quotes,** with position and typography, plus "no extra text". Spell unusual words letter by letter. Use medium or high quality for small text.
6. **Edits: separate the change from what stays.** "Change only X. Keep the face, layout, lighting and labels."
7. **Number your reference images** and give each a role.
8. **One change per turn,** and feed the last output back. If a region must stay pixel-identical, composite it instead.

## GPT Image prompt examples

The first five are from OpenAI's guide, word for word; the last two are ours.

Photoreal portrait (1024×1536, medium):

Open: https://superduperai.co/library/style-lock-block/

```text
Create a photorealistic candid photograph of an elderly sailor standing on a small fishing boat. He has weathered skin with visible wrinkles, pores, and sun texture, and a few faded traditional sailor tattoos on his arms. He is calmly adjusting a net while his dog sits nearby on the deck. Shot like a 35mm film photograph, medium close-up at eye level, using a 50mm lens. Soft coastal daylight, shallow depth of field, subtle film grain, natural color balance. The image should feel honest and unposed, with real skin texture, worn materials, and everyday detail. No glamorization, no heavy retouching.
```

Ad with an exact tagline:

```text
Give me a cool in culture ad / fashion shot for a brand called Thread. It's a hip young street brand. The ad shows a group of friends hanging out together with the tagline "Yours to Create." Make it feel like a polished campaign image for a youth streetwear audience: stylish, contemporary, energetic, and tasteful. Use clean composition, strong color direction, natural poses, and premium fashion photography cues. Render the tagline exactly once, clearly and legibly, integrated into the ad layout. No extra text, no watermarks, no unrelated logos.
```

Logo on a transparent background (`background="transparent"`, PNG):

```text
Create an original, non-infringing logo for a company called Field & Flour, a local bakery. The logo should feel warm, simple, and timeless. Use clean, vector-like shapes, a strong silhouette, and balanced negative space. Favor simplicity over detail so it reads clearly at small and large sizes. Flat design, minimal strokes, no gradients unless essential. Fully transparent background. Deliver a single centered logo with generous padding, clean alpha edges, and no solid backdrop, scenery, checkerboard, or watermark.
```

App mockup:

```text
Create a realistic mobile app UI mockup for a local farmers market. Show today's market with a simple header, a short list of vendors with small photos and categories, a small "Today's specials" section, and basic information for location and hours. Design it to be practical, and easy to use. White background, subtle natural accent colors, clear typography, and minimal decoration. It should look like a real, well-designed, beautiful app for a small local market. Place the UI mockup in an iPhone frame.
```

Remove an object (attach the photo):

```text
Remove the flower from man's hand. Do not change anything else.
```

Product shot with a label (1536×1024, high):

Open: https://superduperai.co/library/product-hero-shot-generation/

```text
Studio photo of a jar of honey on a wooden table, soft morning light from a window on the left, light haze, the label reads "Linden Honey" in a large serif font. Blurred background, warm tones. No other text, no watermarks.
```

Room edit with constraints (attach the photo):

```text
Change only the sofa to a dark green velvet sofa of the same size and position. Keep the walls, floor, lighting, window and every other object exactly as they are.
```

To turn a still into a clip, see the [first-frame method](/library/image-to-video-first-frame-method/) and our [AI video generation](/products/video/).

## GPT Image vs other image models

On the Artificial Analysis text-to-image arena (read 2 November 2026), OpenAI holds the top three: GPT Image 2.5 Sunburst, 2.5 Flare, then GPT Image 2. Google's Nano Banana 2.1 is #4 at roughly one sixth of the price per image ($0.0336 at 1K). gpt-image-1 is down at #42. Google's previous model, Imagen 4, was shut down in August 2026; see [Imagen 4](/blog/google-imagen-4/).

## FAQ


### Is gpt-image-1 still available?

No. OpenAI shut down gpt-image-1 in the API on 23 October 2026. Use gpt-image-2 or gpt-image-2.5-flare instead. gpt-image-1-mini and gpt-image-1.5 shut down on 1 December 2026.

### What was the maximum resolution of gpt-image-1?

1536×1024 or 1024×1536. gpt-image-1 supported only 1024×1024, 1024×1536 and 1536×1024. It never made 4096×4096. GPT Image 2 and 2.5 accept custom sizes up to 3840 px on the long side.

### Can GPT Image edit an uploaded image?

Yes, and gpt-image-1 could too. The edit endpoint takes several input images and an optional mask; in ChatGPT you attach a photo and describe the change.

### Is GPT Image free?

In ChatGPT, yes: Images 2.5 is available on the free plan with a lower, unpublished limit. The API is paid, from about $0.005 per low-quality image on GPT Image 2.

### What is the difference between GPT Image 2.5 Sunburst and Flare?

Sunburst is the larger model, aimed at the highest image quality. Flare is smaller and faster, with quality OpenAI describes as comparable to GPT Image 2. Both cost the same per token.


## Sources

Checked on 2 November 2026:

1. [OpenAI: gpt-image-1 model page](https://developers.openai.com/api/docs/models/gpt-image-1) (status: shut down)
2. [OpenAI: deprecations](https://developers.openai.com/api/docs/deprecations)
3. [OpenAI: API changelog](https://developers.openai.com/api/docs/changelog)
4. [OpenAI: image generation guide](https://developers.openai.com/api/docs/guides/image-generation) (sizes, quality, editing, per-image prices)
5. [OpenAI: image prompting guide](https://developers.openai.com/api/docs/guides/image-prompting)
6. [OpenAI: GPT Image 2](https://developers.openai.com/api/docs/models/gpt-image-2), [2.5 Sunburst](https://developers.openai.com/api/docs/models/gpt-image-2.5-sunburst), [2.5 Flare](https://developers.openai.com/api/docs/models/gpt-image-2.5-flare)
7. [Implicator: ChatGPT Images 2.5](https://www.implicator.ai/openai-ships-chatgpt-images-2-5-with-sketch-tool-and-up-to-50-faster-generation/)
8. [Artificial Analysis: text-to-image](https://artificialanalysis.ai/image/leaderboard/text-to-image)
