GPT-Image 2.6 Flare API
AI ImageGPT Image 2.6 Flare API — Text-to-image generation and reference editing, priced per image in 1K, 2K and 4K tiers like 2.5. Quality does not change the price.
- Per successful call$0.025–$0.055
- OperationsText to image
- Image edit
- View tool pricing and billing rules
Your result will appear here once you run the model.
01
Where it fits
GPT Image 2.6 Flare API — Text-to-image generation and reference editing, priced per image in 1K, 2K and 4K tiers like 2.5. Quality does not change the price.
Make Flare the default 2.6 route: it takes the same size, quality and parameters as 2.5, so moving an existing prompt is a model id change, and at 2048x1152 high it returned in 22 seconds.
02
What it does
Text to image
gpt-image-2.6-flare
Turn a text prompt into a new image.
Image edit
gpt-image-2.6-flare
Change an existing image from a prompt and reference input.
03
In depth
GPT-Image 2.6 Flare is OpenAI's newest image generation and editing id on OmniaKey, published on 4 November 2026 beside GPT-Image 2.6 Sunburst. Both take a text prompt or a prompt plus a reference image, both are listed for text-to-image and image-edit, and both sit on the same per-call resolution tiers as the 2.5 ids.
Flare is the everyday route of the pair. Our first run on 5 November 2026 behaved like 2.5: a 2048x1152 frame at high quality came back in 22 seconds, a reference edit returned a valid PNG, and the request schema is unchanged, so moving a script from 2.5 is a model-id change rather than a rewrite.
The first frame of a new brief
Use Flare while the prompt is still moving. It accepts the same size and quality values as 2.5, so an existing playground or script only needs the new id, and the returned canvas shows whether the composition is worth keeping before you spend on a larger size.
Product and scene photography
Ask for one object or a built set at an explicit size such as 2048x1152 with high quality. Read the whole frame, corners included, and keep a second prompt ready that names the props you want removed; unrequested objects are the failure we see most on this family.
Reference edits
Send the source image inline, set operation to edit, and ask for the canvas you will crop from. A square request is normalised, so compare the returned file against the source side by side rather than assuming the size you sent came back.
04
Hands-on notes
In our 5 November 2026 run, GPT-Image 2.6 Flare returned a 2048x1152 PNG at high quality in 22 seconds.
Read the testThe pricing feed listed the same three resolution tiers for the 2.5 and 2.6 ids, and a 1K edit billed at the 1K tier.
That edit came back as a 1254x1254 PNG after 34 seconds, reporting 416 input tokens and 2058 output image tokens.
05
Limits to plan around
- Generation reports no token usage on 2.6: both generate tasks returned zero counters, so a usage dashboard shows nothing for them and the flat resolution price is the only charge.
- The square canvas you send is not the square you get. A 1024x1024 request, like an omitted size, is normalised to 1254x1254 on the GPT Image routes.
- 4K stays a larger canvas rather than more detail, and it is where upstream timeouts appeared on the 2.5 ids; budget a retry before you promise a 4K delivery.
- SVG output, animated GIF output, adjustable complexity and markup-guided editing are 2.6 features upstream that this API does not document yet, so do not plan a pipeline around them.
06
Compare nearby models
| Model | Best for | Operations | Listed price |
|---|---|---|---|
| GPT-Image 2.6 FlareThis page gpt-image-2.6-flare | New work on the newest generation, where the 2.5 parameters and tiers already fit the pipeline. | Text to image · Image edit | $0.025–$0.055 per call |
| GPT-Image 2.5 Flare gpt-image-2.5-flare | The previous generation with the same contract, for a brief already tuned on 2.5 or an id with a longer track record. | Text to image · Image edit | $0.025–$0.055 per call |
| GPT-Image 2.6 Sunburst gpt-image-2.6-sunburst | Edits and layouts that depend on precise changes, paired with Flare in the same generation. | Text to image · Image edit | $0.025–$0.055 per call |
07
Sources and verification
Guide updated 2026-11-05.
08
Call it
Media models run asynchronously. Create a task, poll it until it finishes, then download the signed asset URL.
# 1. Create the job. 202 Accepted; the body carries the task id,
# the pre-charged price, and a poll_url. Reuse one Idempotency-Key
# across retries so a timeout never starts a second paid job.
curl --fail-with-body https://api.omniakey.com/v1/media/tasks \
-H "Authorization: Bearer $OMNIAKEY_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: gpt-image-2.6-flare-001" \
-d '{
"model": "gpt-image-2.6-flare",
"type": "image",
"operation": "generate",
"input": {"prompt": "a paper boat on a quiet lake at sunset"},
"parameters": {"size": "1024x1024", "quality": "medium"}
}'
# 2. Poll GET /v1/media/tasks/{id} until status is
# succeeded | failed | canceled. Progress carries percent and stage.
curl --fail-with-body https://api.omniakey.com/v1/media/tasks/TASK_ID \
-H "Authorization: Bearer $OMNIAKEY_API_KEY"
# 3. Download assets[0].url before expires_at, or poll again for a
# fresh signed URL. Failed and canceled jobs are refunded.
curl -sS -o out.png "$(jq -r '.assets[0].url' result.json)"09
Pricing
Image jobs are billed per successful call, not per token. Failed or canceled jobs are refunded automatically.
$0.025–$0.055per callSave 53%
- 1K
- $0.025per call
- 2K
- $0.035per call
- 4K
- $0.055per call
More OpenAI image models
Related articles
GPT Image 2.6 vs GPT Image 2.5: What Changes?
OpenAI released GPT Image 2.6 on November 4: faster, with SVG and GIF output, but up to 4.2x the tokens of 2.5 per image.
Read articleGPT Image 2.5 API: Generate and Edit Images in Python
One POST creates the task and one GET reads it back. Here is the request code, the sizes we ran on October 29, 2026, and one edit shown before and after.
Read articleGPT Image 2.5 vs Nano Banana 2.1: Same-Prompt Test and Price
Four prompts, three image models, one day: what each made and what it really cost.
Read articleSeedream 5.0 Pro vs GPT Image 2: Same-Prompt Test and Price
Six jobs, the same prompts, three models: what each made and what it cost.
Read articleFrequently asked questions
How do I call GPT-Image 2.6 Flare on OmniaKey?
POST to /v1/media/tasks with the model id, type "image", and operation "generate" or "edit". Reuse one Idempotency-Key across retries so a timeout never starts a second paid job.
What does GPT-Image 2.6 Flare cost?
GPT-Image 2.6 Flare bills $0.025–$0.055 per call. Creation pre-charges that amount from your prepaid balance, and failed or canceled jobs are refunded automatically.
Why does the API return a task id instead of the image?
Provider latency runs from seconds to about a minute, so OmniaKey normalizes every image route into one asynchronous lifecycle. Poll GET /v1/media/tasks/{id} until the status is succeeded, then read the signed asset URL.
Is GPT-Image 2.6 more expensive than 2.5?
No. The catalog lists the same three resolution tiers for both generations, and the pricing feed returned identical rows for the four 2.5 and 2.6 ids on 5 November 2026. Quality does not change the charge, and generation and editing use the same tier.
How do I call GPT-Image 2.6 Flare?
Create a task with POST /v1/media/tasks, model gpt-image-2.6-flare and an Idempotency-Key header, then poll GET /v1/media/tasks/{id} until the state reads succeeded and download assets[].url. Put size and quality in parameters; for an edit, set operation to edit and pass the source image inline.
Which quality values work?
low, medium and high. As on 2.5, the documented xhigh and max values are accepted and then rendered as medium, so naming a higher tier does not buy more detail. Check the file you receive instead of trusting the parameter.
Can I ask for a square 1024x1024 image?
You can, but the GPT Image routes normalise that square, and an omitted size, to 1254x1254. Request the canvas you need, confirm the dimensions in the downloaded file, and crop from there.
