GPT Image 2.5 is live — OpenAI's newest image model, targeted edits that leave the rest of the frame alone

GPT Image 2.6 — Faster Images, GIFs and SVG

GPT Image 2.6 is OpenAI's image model released on November 4, 2026, with two API tiers: Flare for speed and Sunburst for detailed edits. GPT Image 2.6 adds animated GIF and SVG output and a complexity setting. It is not callable on reAPI yet: until it opens, the playground on this page runs GPT Image 2.5 on the same endpoint.

Text to imageImage editingAnimated GIFSVG outputComplexity control
Highlights
GPT Image 2.6 ships two API tiers, Flare and Sunburst · animated GIF and SVG output · a new complexity setting
Best for
Fast drafts, animated loops, vector logos and icons, consistent edit rounds
Input
Text prompt, input images, optional markup layer
Output
png, jpeg or webp, plus svg and animated gif
Input
Channel
Model

GPT Image 2.6 is not callable on reAPI yet. Until it opens, every request from this playground runs GPT Image 2.5 and is billed at the GPT Image 2.5 price.

required · up to 32,000 characters

0/16 files

optional · up to 16 images

0/1 file

optional · requires at least one reference image

auto or a WIDTHxHEIGHT value

auto · low to max · affects price

auto, opaque or transparent

png, jpeg or webp

100

0–100 · jpeg and webp only

auto or low · website fixed at auto

Estimated cost~$0.02121 credits
Result

Try one of these prompts

Pricing

Credit-based — 1 credit = $0.001 USD. Pay only for completed generations.

PRICE EST.gpt-image-2.5-flare-official
× 1 · real-time

Figures are pre-submit estimates. The final charge uses the actual text-input, image-input, and image-output usage returned by the model.

Your estimate
…
Per $10
… images
medium quality · 1024×1024
Official
~$0.01314 cr

Reference calculator estimate. For batches it multiplies the per-image subtotal; final billing uses the provider's aggregate usage and does not multiply shared inputs again.

Tokens per image
image output439
image input0
text input12
Quality
Size
Pixels
×

mult. 16 · 655K–8.29M px · edges ≤ 3840 · aspect ≤ 3:1

Count
1
With reference images (image-to-image)
0
Prompt
0 chars · ~0 tokens
First test cost
$0.02121 credits

One sample at the cheapest tier (1 image · standard · auto / low / medium).

Testing budget guide
Add credits
$10
≈ 476 tests
$50
≈ 2380 tests
$100
≈ 4761 tests
Category
GPT Image 2.5 Flare · Standard
GPT Image 2.5 Sunburst · Standard
auto / low / medium
$0.021
Price · 1 image · standard · auto / low / medium
$0.021
Price · 1 image · standard · auto / low / medium
high
$0.021
Price · 1 image · standard · high
$0.021
Price · 1 image · standard · high
xhigh
$0.026
Price · 1 image · standard · xhigh
$0.026
Price · 1 image · standard · xhigh
max
$0.026
Price · 1 image · standard · max
$0.026
Price · 1 image · standard · max
GPT Image 2.5 · Official
Metered per token, by quality and size
$0.002
Price · 1 image · official, from

What is GPT Image 2.6?

GPT Image 2.6 is the image model OpenAI released on November 4, 2026, alongside Images 2.6 in ChatGPT and Codex. In the API it ships as two models. GPT Image 2.6 Flare is built for fast iteration and, per OpenAI, can return an image in as little as six seconds at low quality. GPT Image 2.6 Sunburst is built for detailed work and keeps characters, products and brand details more consistent across edits at medium or high quality. OpenAI says Images 2.6 generates images 50% faster than before. Both GPT Image 2.6 models take text and image input and add three things to the Image API: animated GIF output, SVG output, and a complexity setting that controls how much the model plans before rendering. reAPI has not opened GPT Image 2.6 yet. This page runs GPT Image 2.5 in the meantime, on the endpoint and key every reAPI image model uses.

What you can build with this model

Real-world workflows and production use cases you can build and ship with this model.

GPT Image 2.6 turning a brand mascot into a looping animated GIF

Turn a mascot into a looping GIF

GPT Image 2.6 can return an animated GIF straight from a prompt: describe the subject, the motion and the loop. OpenAI pitches it for brand mascots, reactions and memes, and animated storyboards before video production. With GPT Image 2.6, frame count and frame timing are set per request.

GPT Image 2.6 producing a logo as an editable SVG vector file

Get a logo as editable SVG

Ask GPT Image 2.6 for SVG and you get vector paths, shapes and colors instead of pixels, ready for websites, apps and print. It suits logos, icons and flat illustrations. SVG output from GPT Image 2.6 is for new generations, not edits.

Fast GPT Image 2.6 Flare drafts with one picked and refined on Sunburst

Draft fast, then refine the keeper

Explore layouts on GPT Image 2.6 Flare, which OpenAI says can finish an image in about six seconds at low quality. Move the chosen direction to GPT Image 2.6 Sunburst at medium or high quality when it has to stay consistent through further edits.

What GPT Image 2.6 changes

Faster on every round

OpenAI says GPT Image 2.6 generates images 50% faster than before. GPT Image 2.6 Flare also improves layout quality for posters, product imagery and other professional visuals.

Edits that hold together

GPT Image 2.6 Sunburst keeps characters, products and brand details more consistent from one edit to the next when quality is set to medium or high.

Control how much it plans

The complexity setting tells GPT Image 2.6 how much planning a request gets: low for quick edits, medium or high for dense scenes with many constraints. It has no separate fee.

GPT Image 2.6 vs GPT Image 2.5

GPT Image 2.5 is the generation GPT Image 2.6 follows, and it is the model this page runs on reAPI today. Every row below comes from OpenAI's own release post and image generation guide.

Capability
GPT Image 2.6
GPT Image 2.5
Speed
50% faster than before, per OpenAI; Flare can finish in about six seconds at low quality
Flare was described at launch as roughly half the latency of GPT Image 2
Animated GIF output
Yes: 2 to 64 frames, 20 to 2,000 ms per frame
Not listed for GPT Image 2.5 in OpenAI's guide
SVG output
Yes, on image generation requests
Not listed for GPT Image 2.5 in OpenAI's guide
Complexity setting
Low, medium or high planning effort, high by default, no separate fee
Not available; OpenAI introduces it with 2.6
On reAPI
GPT Image 2.6 is not callable on reAPI yet
Live on Standard and Official channels

Comparison reflects publicly documented behavior at the time of writing, drawn from OpenAI's November 4, 2026 release post and its image generation guide. Markup-guided edits are supported on both models and are not listed as a difference.

Get ready for GPT Image 2.6 in three steps

  1. step 01

    Create an API key

    Sign up at reAPI and create a key from the dashboard. One key reaches every model on the platform, so there is no separate GPT Image 2.6 credential.

    Open
  2. step 02

    Build on GPT Image 2.5 today

    POST to /api/v1/images/generations with gpt-image-2.5-flare or gpt-image-2.5-sunburst and poll the returned task id. Every reAPI image model uses this endpoint.

    Open
  3. step 03

    Switch when 2.6 opens

    When GPT Image 2.6 opens on reAPI, this page will list its model ids. Moving over is a change of model id, plus the new GIF, SVG and complexity fields if you want them.

    Open
docs/api/gpt-image-2-6

API reference

Drop-in code and the full parameter table.

curl -X POST https://reapi.ai/api/v1/images/generations \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-image-2.5-flare",
    "prompt": "An editorial poster for a ceramic kettle, soft window light",
    "size": "1536x1024",
    "quality": "high"
  }'

Parameter validation failed

Rejected synchronously with the failing field named. Check enum values, ranges, and URL fields — nothing is charged.

Authentication or balance issue

401 means a missing or invalid Bearer key; 402 means the reserve exceeds your balance. Manage keys and credits in the console.

Content or material rejected

Reference material and outputs pass automated moderation. Rejected tasks fail with a clear error and a full refund.

Task failed or timed out

A task that reaches a failed state is never charged — the reserve refunds automatically. Keep the task id and retry when ready.

Frequently asked questions

Common questions about this model.

They are the two GPT Image 2.6 models in OpenAI's API, and both support complexity control, markup-guided edits, SVG and animated GIF output. Flare is for fast iteration: at low quality it can return an image in as little as six seconds, and it improves layout quality for posters and product imagery. Sunburst is for detailed visual work: at medium or high quality it gives higher-quality generation and keeps details more consistent across edits.

start building

Ready to ship?

Try it in the playground or grab an API key to integrate now.