GPT Image
GPT Image is OpenAI's family of image generation and editing models in the API, from gpt-image-1 (Apr 2025) to GPT-Image-2.6 Sunburst and Flare (Nov 4, 2026). The current models cost $8 per 1M image input tokens and $30 per 1M image output tokens, and power ChatGPT Images.
GPT Image is the name of OpenAI's image models in the API, and as of Nov 6, 2026 the newest are GPT-Image-2.6 Sunburst (gpt-image-2.6-sunburst, "for detailed visual work") and GPT-Image-2.6 Flare (gpt-image-2.6-flare, "for fast iteration"), released on Nov 4, 2026 with SVG and animated GIF output. Both cost $8 per 1M image input tokens and $30 per 1M image output tokens, plus $5 per 1M text input tokens. The same models power ChatGPT Images for ChatGPT users. The line began with gpt-image-1 on Apr 23, 2025, which replaced DALL-E and was itself shut down on Oct 23, 2026. For the rest of the platform, see OpenAI API.
Quick facts
| Newest models | gpt-image-2.6-sunburst and gpt-image-2.6-flare, Nov 4, 2026 |
| Default for most apps | GPT-Image-2.5 Flare, per OpenAI's Sep 8, 2026 post |
| Price (2.6, 2.5, 2) | $8 image in, $2 cached, $30 image out; $5 text in per 1M tokens |
| Per image (GPT Image 2, 1024x1024) | About $0.006 low, $0.053 medium, $0.211 high |
| Endpoints | v1/images/generations, v1/images/edits, and the image tool in Responses |
| Formats added in 2.6 | SVG and animated GIF |
| Largest edge | 3,840 px; above 2560x1440 is experimental |
| Weekly images | 4 billion across ChatGPT, Codex and the API, per OpenAI on Nov 4, 2026 |
Every GPT Image model
| Model | Released | Status on Nov 6, 2026 |
|---|---|---|
| gpt-image-1 | Apr 23, 2025 | Shut down Oct 23, 2026 |
| gpt-image-1-mini | Oct 6, 2025 | Shuts down Dec 1, 2026 |
| gpt-image-1.5 (and chatgpt-image-latest) | Dec 16, 2025 | Shuts down Dec 1, 2026 |
| gpt-image-2 | Apr 21, 2026 | Current |
| gpt-image-2.5-sunburst, gpt-image-2.5-flare | Sep 8, 2026 | Current |
| gpt-image-2.6-sunburst, gpt-image-2.6-flare | Nov 4, 2026 | Current |
OpenAI names GPT-Image-2.6 Sunburst or Flare as the replacement for everything that shuts down on Dec 1, 2026 (deprecations).
Sunburst or Flare?
Since version 2.5, each release comes in two models. Sunburst is the quality model and Flare the fast one. When 2.5 launched, OpenAI called Flare "the default choice for most applications", with higher quality than GPT-Image-2 at 50% lower latency, and Sunburst the model "built for premium visual workflows" (OpenAI, Sep 8, 2026); OpenAI's API video the same day called Flare "over 50% faster than GPT Image 2" at "equal quality". With 2.6, OpenAI says images generate 50% faster than before, and that Flare at low quality can return an image "in as little as six seconds" (OpenAI, Nov 4, 2026). Both are OpenAI's own figures.
The two cost the same per token (pricing), so the choice is about speed and detail rather than the rate card. Per image, though, OpenAI warns that "equal token rates don't mean equal cost per image" (image generation guide), because models use different numbers of output tokens.
What GPT-Image-2.6 adds
From OpenAI's image generation guide, read Nov 6, 2026 (guide):
complexity:low,mediumorhigh(defaulthigh), with "no separate fee".- SVG output:
output_format: "svg", one image per request (n= 1), generation only, not edits. - Animated GIF output:
output_format: "gif", withframe_countfrom 2 to 16 (default 9) andframe_duration_msfrom 20 to 2,000 (default 200). - Custom sizes: multiples of 16, aspect ratios from 1:3 to 3:1, edges up to 3,840 px, and 655,360 to 8,294,400 pixels in total. Sizes above 2560x1440 are "experimental".
- Quality: the 2.5 models accept
low,medium,high,xhigh,maxandauto(GPT-Image-2.5 Sunburst); GPT Image 2 acceptslow,mediumandhigh(gpt-image-2).
OpenAI's Nov 4, 2026 launch video also shows markup editing: you send a separate image with a circled area or annotations, describe the change, and the model makes that specific edit.
One oddity: the guide's overview still names the 2.5 models as the ones it documents, even though it describes the 2.6 features (guide).
How much does GPT Image cost?
| Model | Image input / cached / output | Text input / cached | Batch |
|---|---|---|---|
| gpt-image-2.6-sunburst, -flare | $8.00 / $2.00 / $30.00 | $5.00 / $1.25 | No Batch row listed |
| gpt-image-2.5-sunburst, -flare | $8.00 / $2.00 / $30.00 | $5.00 / $1.25 | Half price |
| gpt-image-2 | $8.00 / $2.00 / $30.00 | $5.00 / $1.25 | Half price ($4.00 / $1.00 / $15.00) |
| gpt-image-1.5, chatgpt-image-latest | $8.00 / $2.00 / $32.00 | $5.00 / $1.25 | |
| gpt-image-1-mini | $2.50 / $0.25 / $8.00 | $2.00 / $0.20 |
Source: OpenAI API pricing, per 1M tokens, read Nov 6, 2026. Cached input for GPT Image 2 and 2.5 applies only to images generated through the Responses API.
Per image. The guide lists fixed per-image costs only for the older models. For a 1024x1024 square: GPT Image 2 about $0.006 (low), $0.053 (medium) and $0.211 (high); GPT Image 1.5 $0.009, $0.034 and $0.133; GPT Image 1 Mini $0.005, $0.011 and $0.036. Portrait and landscape cost differently. The guide offers a calculator rather than a table for 2.5 and 2.6, and each partial image streamed along the way adds 100 output tokens.
How to generate an image with the API
- Install the SDK and set
OPENAI_API_KEY(see OpenAI API). - Send a prompt to
v1/images/generationswithmodelset togpt-image-2.6-flarefor speed orgpt-image-2.6-sunburstfor detail (model page). - Choose
quality, a size, and anoutput_format(for examplesvgorgifon 2.6). - To change an existing image, send it with a prompt to
v1/images/edits. - Inside a conversation, use the image generation tool of the Responses API instead, which also makes cached image input possible on GPT Image 2 and 2.5.
For bulk jobs, the Batch API accepts /v1/images/generations and /v1/images/edits at half price for models that have a Batch row (Batch guide).
Provenance and rules
Images from the first GPT-4o image model carry C2PA metadata that identifies them as AI-generated (OpenAI, Mar 25, 2025), and gpt-image-1 kept C2PA metadata and added a moderation parameter (OpenAI, Apr 23, 2025). The pages read for this article do not restate the C2PA policy for 2.5 and 2.6. OpenAI published a system card for ChatGPT Images 2.6 on Nov 4, 2026.
How GPT Image compares
Google's Nano Banana models are the main rival in the API: gemini-nano-banana-2.1 costs $0.0336 per image at 1K resolution, against about $0.053 for a medium GPT Image 2 square, though the two quality levels are not equivalent. ChatGPT Images vs Nano Banana compares them properly. The image arena leaderboards were not read for this page, so no quality ranking is given here.
Frequently asked questions
What is the latest OpenAI image model? As of Nov 6, 2026, the latest are GPT-Image-2.6 Sunburst and GPT-Image-2.6 Flare, released Nov 4, 2026 with SVG and animated GIF output; Sunburst targets detail and Flare speed.
How much does gpt-image-2 cost per image? OpenAI's guide puts a 1024x1024 GPT Image 2 image at about $0.006 in low quality, $0.053 in medium and $0.211 in high, from token prices of $8 per 1M image input tokens and $30 per 1M image output tokens.
Is gpt-image-1 still available? No. OpenAI shut gpt-image-1 down on Oct 23, 2026; gpt-image-1.5, gpt-image-1-mini and chatgpt-image-latest follow on Dec 1, 2026, with GPT-Image-2.6 as the replacement.
Can the OpenAI image API make GIFs and SVGs? Yes, with GPT-Image-2.6. Set output_format to gif for an animated GIF of 2 to 16 frames, or to svg for a vector image, one per request.
Does OpenAI have a video API? Not anymore. OpenAI shut down the Sora 2 models and its Videos API on Sep 24, 2026 with no replacement, so GPT Image covers only still and animated GIF images; see Sora.
Sources
- Image generation, OpenAI, read Nov 6, 2026
- GPT-Image-2.6 Sunburst model page, OpenAI, read Nov 6, 2026
- GPT-Image-2.5 Sunburst model page, OpenAI, read Nov 6, 2026
- gpt-image-2 model page, OpenAI, read Nov 6, 2026
- gpt-image-1 model page, OpenAI, read Nov 6, 2026
- Pricing, OpenAI, read Nov 6, 2026
- Batch API, OpenAI, read Nov 6, 2026
- Changelog, OpenAI, read Nov 6, 2026
- Deprecations, OpenAI, read Nov 6, 2026
- Introducing Images 2.6, OpenAI, Nov 4, 2026
- Introducing ChatGPT Images 2.5, OpenAI, Sep 8, 2026
- Introducing ChatGPT Images 2.0, OpenAI, Apr 21, 2026
- The new ChatGPT Images is here, OpenAI, Dec 16, 2025
- Image generation API, OpenAI, Apr 23, 2025
- Introducing 4o Image Generation, OpenAI, Mar 25, 2025
- Deployment Safety Hub, OpenAI, read Nov 6, 2026
- Gemini API pricing, Google, read Nov 6, 2026
- GPT-Image-2.6: Faster image generation with Flare and Sunburst, OpenAI, Nov 4, 2026 (video)
- Introducing GPT-Image-2.5 in the API, OpenAI, Sep 8, 2026 (video)
Moments
GPT-Image-2.6: Faster image generation with Flare and Sunburst
OpenAI
Introducing GPT-Image-2.5 in the API
OpenAI
Photos
See also
-
ChatGPT ImagesChatGPT Images makes and edits images from a prompt in any chat, on every plan; the current version, Images 2.6 since Nov 4, 2026, adds looping GIFs and saved reference images. Free gets limited images, Plus and Pro get images with thinking, and Pro is unlimited within abuse guardrails. - DALL-EDALL-E was OpenAI's text-to-image model line: DALL-E (Jan 2021), DALL-E 2 (2022) and DALL-E 3 (2023, built into ChatGPT). GPT-4o image generation replaced it in ChatGPT in March 2025, the API models shut down on May 12, 2026, and the DALL-E GPT retired on Aug 30, 2026.
-
OpenAI APIThe OpenAI API lets programs call OpenAI's models (GPT-6 Astra, GPT-6.1 Sol, GPT-6.1 Terra, GPT-6 Luna, image, voice and transcription models) with an API key, billed per token. New projects use the Responses API through the official SDKs, with documentation at developers.openai.com.
-
Responses APIThe Responses API is OpenAI's recommended endpoint for new projects, combining model calls, built-in tools, structured outputs and conversation state in one request. Chat Completions is still supported, and the Assistants API it replaced shut down on Aug 26, 2026.
-
Batch APIThe OpenAI Batch API runs up to 50,000 requests from one file within 24 hours at 50% off the normal price. Flex processing gives the same discount on ordinary calls that can be slow, Fast mode costs double for quicker answers, and prompt caching cuts the price of repeated prompt prefixes.
-
ChatGPT Images vs Nano BananaAs of Nov 6, 2026 OpenAI's GPT Image models behind ChatGPT Images hold the top five places on Arena's text-to-image leaderboard, ahead of Google's Nano Banana 2.1 in seventh, while Nano Banana is cheaper per image in the API; Midjourney is not on that leaderboard and remains the choice for its own look, and FLUX is the cheapest high-volume option.