GPT Image 2.5: leave quality blank and you pay 7 cents, not 1
On Sume, GPT Image 2.5 defaults to high quality when you omit the field. A 1024-class image is 7 cents at high, 2 at medium and 1 at low. 100 images: $7 vs $1.
The short answer
If you call GPT Image 2.5 on Sume and leave out quality, the model runs at high. The Sume images page says the default is high for both ChatGPT Image 2.5 variants. At a 1024-class size that is 7 cents per image, against 2 cents at medium and 1 cent at low. Over 100 images the unset default costs $7.00, and the low tier costs $1.00.
Sume's billable amount is the provider list price times 1.25, rounded up to the cent. The per-image figures below are for 1024x1024; other sizes and image-input tokens move the high tier within roughly 6 to 7 cents.
Price per image at each GPT Image 2.5 tier
Quality accepts auto, low, medium, high, xhigh or max. The last two are priced from the token calculator described in the docs.
| Quality | Per image | 100 images | Notes |
|---|---|---|---|
| low | $0.01 | $1.00 | Cheapest tier; use for drafts |
| medium | $0.02 | $2.00 | Between draft and final |
| high (the default) | $0.07 | $7.00 | Catalog list $0.0527 x 1.25 = $0.065875, rounded up |
| xhigh | $0.12 | $12.00 | Docs: $0.09366 output, rounded up to $0.0937; x 1.25 = $0.117125 |
| max | $0.27 | $27.00 | Docs: $0.21072 output, rounded up to $0.2108; x 1.25 = $0.2635 |
The arithmetic
The images page states the Fal token rates for both Flare and Sunburst: $30 per million output image tokens, $8 per million input image tokens and $5 per million input text tokens. At 1024x1024 it gives $0.09366 for xhigh and $0.21072 for max, before input tokens and Sume pricing. Fal rounds the total up to $0.0001.
For high, the Sume catalog shows $0.0527 per image (the high, 1024 display rate). Times 1.25 that is $0.065875, which rounds up to 7 cents. For xhigh: $0.0937 x 1.25 = $0.117125, which rounds up to 12 cents. For max: $0.2108 x 1.25 = $0.2635, which rounds up to 27 cents.
If you send reference images or long prompts, input tokens are added before the 1.25 factor, so the high tier can land at 6 or 7 cents depending on the request.
What to do about it
Set quality explicitly on every call. A draft loop at low costs 1 cent per image; a promoted final at high costs 7. Two more points from the docs matter for budgets. quality: auto reserves the max rate while the request is admitted, so a wallet needs room for 27 cents per image on those calls. And generation is billed all-or-nothing: a failed or cancelled generation is not charged.
See Images for the request fields, including num_images (1 to 4), which multiplies the per-image price.
Sources
Related posts
More in Pricing
- GPT Image 2.5 quality auto: 26.35 cents reserved vs 6.59 at high
Leave GPT Image 2.5 on quality auto and Sume estimates the max-quality price: $0.2635 for a square image against $0.0659 at high. Pin quality to control it.
- GPT Image 2.5 for packaging text: default high, xhigh and max prices
GPT Image 2.5 on Sume defaults to quality high. At 1024x1024 the provider output is $0.09366 at xhigh and $0.21072 at max, about $0.117 and $0.263 billed.
- gpt-image-2.5 xhigh and max at 2048 and 4K: up to $0.54 an image
gpt-image-2.5 max costs $0.5353 at 2048x2048 and $0.5004 at 3840x2160 on Sume; xhigh is $0.2379 and $0.2224. Size-by-tier table and a spend guard in Python.
- GPT Image 2 or 2.5 on Sume: 27 cents vs 7 cents per high image
At the catalog high/1024 rate, GPT Image 2 is $0.211 list and 27 cents billable on Sume; GPT Image 2.5 is $0.0527 list and 7 cents. 100 images: $27 vs $7.
Written by Sume