GPT Image 2.5 cost per image by quality: the token math on Sume

At 1024x1024, GPT Image 2.5 output costs $0.09366 at xhigh quality and $0.21072 at max, before input tokens and before Sume's pricing, per Sume's Image API docs. The rates behind it are $30 per million output image tokens, $8 per million input image tokens and $5 per million input text tokens.

OpenAI's pricing page lists the same style of rates for its GPT Image models (text input $5.00, image input $8.00, image output $30.00 per 1M tokens for gpt-image-2) and points to a calculator for per-image estimates rather than printing per-image prices.

Working backwards from the docs

Dividing the documented output costs by $30 per million tokens gives the token counts: $0.09366 / $30 x 1,000,000 = 3,122 output tokens at xhigh, and $0.21072 / $30 x 1,000,000 = 7,024 at max. The ratio of the two is 2.25, so max costs 2.25 times xhigh at this size.

Budgeting without surprises

Two reservation rules in the docs matter for a batch. auto quality reserves max, and auto size or named presets without a verified pixel mapping reserve the output-token upper bound. Omitted quality defaults to high, so set quality and a custom size explicitly if you want a tight estimate.

For 100 images at 1024x1024, output alone is $9.37 at xhigh and $21.07 at max before input tokens and Sume pricing; Sume's charged price is shown on the model's endpoint pricing line.