GPT-Image 2.6 Flare & Sunburst are liveImages from $0.025 each

GPT-Image 2.6 Sunburst API

AI Image

GPT Image 2.6 Sunburst 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.

Read the docsGet your API key

OpenAI

GPT-Image 2.6 Sunburst

Mode
Prompt0 / 32000
Quality

Higher quality adds detail. The price stays the same.

Aspect ratio
Resolution

Output 1024 × 1024

1K, 2K, and 4K set the pixel target; the exact size follows the aspect ratio.

Estimated $0.025

Your result will appear here once you run the model.

01

Where it fits

GPT Image 2.6 Sunburst 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.

Use Sunburst for reference edits and final layouts: it is OpenAI's editing-precision variant of the new generation, and it measured the same cost and behaviour as Flare at the same size and quality.

02

What it does

Text to image

gpt-image-2.6-sunburst

Turn a text prompt into a new image.

Image edit

gpt-image-2.6-sunburst

Change an existing image from a prompt and reference input.

03

In depth

GPT-Image 2.6 Sunburst is the second of the two 2.6 ids on OmniaKey, published on 4 November 2026 beside GPT-Image 2.6 Flare. OpenAI keeps the same division of labour it used for 2.5: Sunburst is the id it positions for editing precision, and the catalog gives both ids the same capabilities and the same per-call resolution tiers.

On our first run the pair behaved like 2.5. Sunburst turned a 2048x1152 prompt at high quality around in 26 seconds, and the edit path accepted a reference image and returned a valid PNG. The request schema did not change, so an existing 2.5 pipeline can switch by id alone.

Edits that must land exactly

List what changes and what must not, then send the source inline with operation set to edit. Sunburst is the id OpenAI positions for this work, and the honest check is a side-by-side comparison of the result and the source, including the lettering and the edges.

Layouts a reader will inspect

Menus, price columns and labels reward a prompt that states the column rule and the capitalisation in words. Ask for the exact strings, then proofread the last line before you publish; text-heavy frames are where a single wrong character costs the whole render.

Finals after a Flare draft

Draft the composition on Flare, then repeat the accepted prompt on Sunburst for the frame that ships. Both sit on the same tiers and take the same parameters, so the comparison is a like-for-like test of the two ids on your own brief.

04

Hands-on notes

  • In our 5 November 2026 run, GPT-Image 2.6 Sunburst returned a 2048x1152 PNG at high quality in 26 seconds.

    Read the test
  • Both 2.6 ids were listed at the 2.5 tiers, and an edit on 2.6 accepted a 2048x1152 source and returned a valid PNG.

05

Limits to plan around

  • Generation reports no token usage on 2.6, so a dashboard built on usage counters stays empty for these tasks and the flat resolution price is the whole charge.
  • A 1024x1024 request is normalised to 1254x1254 on this route, so a square brief needs a crop step rather than a smaller request.
  • Quality above high does not exist: xhigh and max are accepted and rendered as medium on this family, which is also why a flat price cannot reflect a higher setting.
  • SVG output, animated GIF output, adjustable complexity and markup-guided editing are upstream 2.6 features this API has not documented or smoked, so treat them as unavailable.

06

Compare nearby models

ModelBest forOperationsListed price
GPT-Image 2.6 SunburstThis page
gpt-image-2.6-sunburst
Precise edits and reader-facing layouts on the newest generation.Text to image · Image edit
$0.025–$0.055
per call
GPT-Image 2.5 Sunburst
gpt-image-2.5-sunburst
The same role in the previous generation, for briefs already tuned there or ids with a longer record.Text to image · Image edit
$0.025–$0.055
per call
GPT-Image 2.6 Flare
gpt-image-2.6-flare
Drafts, product frames and the first pass of a prompt, paired with Sunburst for the final edit.Text to image · Image edit
$0.025–$0.055
per call

08

Call it

Media models run asynchronously. Create a task, poll it until it finishes, then download the signed asset URL.

Create, poll, download
# 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-sunburst-001" \
  -d '{
    "model": "gpt-image-2.6-sunburst",
    "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.

Per successful call

$0.025–$0.055per callSave 53%

1K
$0.025per call
2K
$0.035per call
4K
$0.055per call

More OpenAI image models

Related articles

Frequently asked questions

How do I call GPT-Image 2.6 Sunburst 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 Sunburst cost?

GPT-Image 2.6 Sunburst 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.

How is GPT-Image 2.6 Sunburst different from Flare?

The generation, the parameters and the tiers are the same; the difference is the role OpenAI assigns each id. Sunburst is the editing-precision one, which is why it is the sensible pick when a layout or a reference edit has to come back exact.

Does GPT-Image 2.6 cost more than 2.5?

It does not. The catalog lists the same three resolution tiers for both generations and the pricing feed returned identical rows for the four ids on 5 November 2026, so upgrading is a model-id change, not a new budget line.

Which size should I request?

Send an explicit published size such as 2048x1152 or 1536x1024. Sizes without a configured tier are rejected before the charge, and the GPT Image routes normalise a square 1024x1024 request to 1254x1254, so confirm the dimensions in the downloaded file.

Can I use Batch or a transparent background?

Neither is documented for the 2.x image ids. Transparent backgrounds came back opaque in our 2.5 probes, and Batch is listed for GPT-Image 2 only, so keep background removal and overnight volume on a different route.