ChatGPT Image 2.5 xhigh and max price grid, six sizes on Sume

What xhigh and max cost next to high on openai/gpt-image-2.5, from 1024 square to 4K: $0.117 and $0.264 at 1024x1024, $0.50 for max at 4K, with the default.

5 min readSume
All posts

On Sume, openai/gpt-image-2.5 takes six quality values: auto, low, medium, high, xhigh and max. If you omit quality, the default is high, so a request with only a prompt already costs more than the low rows most price lists lead with. xhigh and max exist only on the 2.5 models (Flare and Sunburst share one price); openai/gpt-image-2 stops at high.

At 1024x1024 the step from high to xhigh is about 1.8x, and from high to max about 4x. The grid below prices every step at six sizes, read 2026-10-07 from Sume's own estimator for the 2.5 token rates (output tokens only, before input tokens for a prompt or references).

Price per image by size and quality

Prices here are Sume's list-times-1.25 figures. The catalog states the billable formula as "list × 1.25 → ceil usd cents", so treat the dollar amounts as the pre-rounding value and read the exact charge from billable_amount_usd_micros in the submit envelope. Failed or cancelled generations are not billed.

Billed price per image, openai/gpt-image-2.5 on Sume, read 2026-10-07. Output-token estimate; input tokens extra.
Sizelowmediumhigh (default)xhighmax
1024x1024$0.00737$0.0165$0.06588$0.11713$0.2635
1280x720$0.004$0.00925$0.03562$0.06313$0.14213
1536x1024$0.006$0.01287$0.0515$0.09225$0.20588
1920x1080$0.00562$0.01287$0.0495$0.08812$0.198
2560x1440$0.00775$0.018$0.06913$0.12287$0.27637
3840x2160$0.014$0.0325$0.12512$0.22237$0.50038

What the numbers say

A 1280x720 max image costs $0.14213, more than a 1920x1080 high image at $0.0495. If the point is a sharper frame, high at a bigger size is cheaper than max at a smaller one.

The cheapest legal path to 4K is low at $0.014; max at 4K is $0.50038. 4K, xhigh and max are also the configurations most likely to outlast the 30-second wait on POST /v1/images, in which case the route answers 202 with a job envelope instead of the image. Branch on the status code, not the body shape.

A practical rule: draft at low, pick the composition, then re-run only the winner at high. Reserve xhigh and max for the one image that will be printed or shown large. Setting quality: "auto" reserves the max price at admission, so do not use it as a cost saver.

The request

Pin the tier explicitly so the default never decides your bill.

curl -X POST https://api.sume.com/v1/images \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"openai/gpt-image-2.5","prompt":"poster of a lighthouse at dusk","quality":"xhigh","image_size":{"width":1920,"height":1088}}'

Sources: Sume Image API docs (read 2026-10-07) and the model catalog at GET /v1/images/models.

Sources

Related posts

More in Pricing

All Pricing posts

Written by Sume