GPT Image
GPT Image is OpenAI's image generation model family, used in ChatGPT (as ChatGPT Images) and in the API. The latest, ChatGPT Images 2.6, arrived on Nov 4, 2026; it is free to try in ChatGPT with limits, and the API version costs $30 per million image output tokens.
GPT Image is OpenAI's family of image models, the engine behind image generation in ChatGPT and the gpt-image models in the OpenAI API; the newest, ChatGPT Images 2.6, arrived on Nov 4, 2026, and in the API it is sold as gpt-image-2.6-flare (fast) and gpt-image-2.6-sunburst (detailed) at $30 per million image output tokens (as of Nov 6, 2026). It is available on every ChatGPT plan, the free one included, with tighter limits on the cheaper plans. Its strengths are following long, specific instructions, writing accurate text inside images and editing by conversation, and on the Arena text-to-image leaderboard (updated Nov 2, 2026) OpenAI models held the top five places.
Use it when you already pay for ChatGPT or want an image that matches a detailed brief (a poster, a diagram, a product mock-up, a comic panel). Look elsewhere for a signature artistic look with little prompting (Midjourney), an open model you can run yourself (Flux, Stable Diffusion), Google's ecosystem and cheaper fast edits (Nano Banana), or video (see Best AI Video Generator). See the Best AI Image Generator for the full comparison.
Learn more: GPT Image on chatgpt.vi.ki
Quick facts
| Maker | OpenAI |
| Launched | In ChatGPT on Mar 25, 2025 (as GPT-4o image generation) |
| Latest | ChatGPT Images 2.6, Nov 4, 2026 |
| Get it | ChatGPT on web, iOS and Android; Codex; the OpenAI API |
| Price | Free in ChatGPT with limits; API from $8 per million image output tokens (gpt-image-1-mini) (as of Nov 6, 2026) |
| API models | gpt-image-2.6-flare, gpt-image-2.6-sunburst, gpt-image-2.5-flare, gpt-image-2.5-sunburst, gpt-image-2, gpt-image-1.5, gpt-image-1-mini |
| Best for | Following detailed prompts, text in images, conversational editing |
| Open weights | No |
Prices and limits
In ChatGPT. Images are on all ChatGPT plans. OpenAI's pricing page describes Free as "limited image creation", Go as more, Plus as "more complex and accurate" images, and Pro as unlimited and faster (subject to abuse guardrails); it does not publish a number of images per day for any plan as of Nov 6, 2026. "Images with thinking", which plans and refines an image before drawing it, is on Plus, Pro and Business, with Enterprise and Edu to follow. Plan prices are on ChatGPT and AI Subscription Prices.
In the API. Billing is by tokens:
| Model | Text input | Image input | Image output |
|---|---|---|---|
| gpt-image-2.6-flare / sunburst | $5 | $8 | $30 |
| gpt-image-2 | $5 | $8 | $30 |
| gpt-image-1.5 | $5 | $8 | $32 |
| gpt-image-1-mini | $2 | $2.50 | $8 |
Prices per million tokens, from OpenAI's pricing page and the gpt-image-2.6-sunburst and gpt-image-2.6-flare model pages, as of Nov 6, 2026. The Batch API halves them. OpenAI no longer lists a fixed price per image; its image guide has a calculator, since cost depends on size and quality.
Features
Instruction following and text. The models read a prompt the way ChatGPT reads a message, so long lists of requirements, exact wording on signs and labels, and layouts such as infographics are their main advantage. With Images 2.0 (Apr 21, 2026), OpenAI added output up to 2K and better text in more languages, and a thinking mode that can look things up on the web and draw a set of up to eight consistent images.
Editing. You can upload a photo or pick a generated one and describe the change, or select an area and describe what goes there. Images 2.6 is, by OpenAI's account, faster and better at keeping details the same from one edit to the next, such as a character's outfit when the background changes.
GIFs, sketches and saved references. Since Nov 4, 2026, ChatGPT can make looping GIFs and animate a still image, and you can save a photo under a name and call it later with @, so a pet or product does not need uploading again. Images 2.5 (Sep 8, 2026) added templates, drawing a sketch to start from, and comment-based edits. The 2.6 API models can also output SVG and animated GIFs.
Provenance. OpenAI's system card for Images 2.0 says generated images carry C2PA metadata and a watermark, and that higher realism raised the deepfake risk its safety layers are meant to block.
History
- Mar 25, 2025: GPT-4o image generation replaced DALL-E as ChatGPT's default. Studio Ghibli-style portraits went viral within days, and on Mar 27 Sam Altman said OpenAI's GPUs were "melting" and limited use while free users waited.
- 2025: The model reached the API as gpt-image-1, then gpt-image-1-mini and gpt-image-1.5.
- Apr 21, 2026: ChatGPT Images 2.0 and gpt-image-2, with thinking.
- Sep 8, 2026: Images 2.5, with the API split into Flare and Sunburst.
- Nov 4, 2026: Images 2.6, GIF creation and saved references.
GPT Image vs the alternatives
On the Arena text-to-image leaderboard (updated Nov 2, 2026, read Nov 6), where people vote between two images blind, GPT-Image-2.5 Sunburst was first with 1427, ahead of Microsoft's MAI-Image-2.6 (1331) and Google's Nano Banana 2.1 (1325). It also led the image edit board. Votes measure preference, not every use, and the newest 2.6 entries had few votes yet.
- Nano Banana: Google's model in Gemini, close behind on edits and with a free tier in a different ecosystem.
- Midjourney: a more consistent artistic style and Personalization, but paid only and with no API.
- Ideogram: a dedicated design tool for typography and layouts.
- Flux and Stable Diffusion: open weights for running locally or fine-tuning, which OpenAI does not offer.
- Adobe Firefly: inside Photoshop and Adobe's apps, and it offers OpenAI's model among its partner models.
Frequently asked questions
Is ChatGPT image generation free? Yes, ChatGPT Images works on the free ChatGPT plan, but with limited image creation; OpenAI does not publish the exact free limit, and paid plans raise it.
What model does ChatGPT use for images? As of Nov 6, 2026, ChatGPT uses ChatGPT Images 2.6, released on Nov 4, 2026; in the API the same generation is gpt-image-2.6-flare and gpt-image-2.6-sunburst.
How much does the GPT Image API cost? OpenAI's gpt-image-2.6 models cost $5 per million text input tokens, $8 per million image input tokens and $30 per million image output tokens, as of Nov 6, 2026; gpt-image-1-mini is cheaper at $8 per million image output tokens.
Is DALL-E still available in ChatGPT? No. OpenAI retired the DALL-E GPT from ChatGPT, and image requests now go to ChatGPT Images.
Can GPT Image edit my photos? Yes. In ChatGPT, upload a photo and describe the change, or select an area with the editor first; the API offers the same through its image edit endpoint.
Sources
- ChatGPT release notes, OpenAI, read Nov 6, 2026
- Images in ChatGPT, OpenAI, read Nov 6, 2026
- ChatGPT pricing, OpenAI, read Nov 6, 2026
- Pricing, OpenAI, read Nov 6, 2026
- gpt-image-2.6-sunburst and gpt-image-2.6-flare, OpenAI, read Nov 6, 2026
- Introducing ChatGPT Images 2.0, OpenAI, Apr 21, 2026
- ChatGPT Images 2.0 System Card, OpenAI, Apr 21, 2026
- Introducing GPT Images 2.5 in the API and ChatGPT, OpenAI Developer Community, Sep 8, 2026
- ChatGPT's viral image-generation AI is 'melting' OpenAI's GPUs, CNBC, Mar 27, 2025
- Text-to-Image and Image Edit leaderboards, Arena, read Nov 6, 2026
- How to Use ChatGPT Images 2.5 (Step-by-Step Tutorial), Kevin Stratvert, Sep 16, 2026 (video)
- Introducing ChatGPT Images 2.6, OpenAI, Nov 4, 2026 (video)
Moments
How to Use ChatGPT Images 2.5 (Step-by-Step Tutorial)
Kevin Stratvert
- GPT Image is OpenAI's family of image models, the engine behind image generation in ChatGPT and the gpt-image models…
- Instruction following and text. The models read a prompt the way ChatGPT reads a message, so long lists of…
- Editing. You can upload a photo or pick a generated one and describe the change, or select an area and describe what…
Introducing ChatGPT Images 2.6
OpenAI
Photos
See also
-
ChatGPTChatGPT is OpenAI's AI assistant. It is free to use, with paid plans of Go ($8 a month in the US), Plus ($20) and Pro ($100, $200 or $500), and since Oct 2026 it runs on the GPT-6 models. -
MidjourneyMidjourney is a paid AI image and video generator from the independent lab Midjourney, Inc. Plans run from $10 to $120 a month with no free tier, and its default model since Jul 24, 2026 is V8.2.
-
FluxFlux is the family of image (and now video) models from Black Forest Labs, some with open weights you can run yourself. The newest open image model, Flux 3 Image Klein (Oct 15, 2026), is free for commercial use under $5 million in revenue, and the API starts at $0.009 an image. -
Stable DiffusionStable Diffusion is Stability AI's family of open image models (SD 1.5, SDXL, SD 3.5) that anyone can download and run on their own computer. SD 3.5 is free to use commercially for those under $1 million in annual revenue, and Stability's API charges from $0.025 an image.
-
Nano BananaNano Banana is Google's name for its Gemini image generation and editing models. The newest, Nano Banana 2.1, came out on Oct 6, 2026; it is free to use with limits in the Gemini app and costs about $0.034 per 1K image in the Gemini API.
-
Best AI Video GeneratorThe best AI video generator in Nov 2026 on LMArena's votes is Google's Gemini Omni Flash, and on Artificial Analysis's it is Alibaba's Wan 3.0; ByteDance's Seedance and MiniMax's Hailuo models rank near the top of both. OpenAI closed the Sora app in 2026, and Sora 2 and Veo 3.1 now rank 13th and 14th on LMArena.