Ideogram 4.5 on Sume: low 4 cents, medium 8, high 28 per image

Ideogram 4.5 costs 4 cents per image at low, 8 cents at medium and 28 cents at high on Sume, at either resolution (1K or 2K). If you omit quality, the default is medium, so a bare request costs 8 cents. High is seven times the price of low. Before a batch of any size, run one prompt at all three and decide on the result.

Where the numbers come from

The Sume Image API page gives the Fal list price as $0.03, $0.06 and $0.22 per image by quality, for all sizes. Sume bills list x 1.25, rounded up to the cent.

What stays the same across tiers

• The text-to-image and edit modes. Without input_references the model generates from text.
• With references it edits the first image and uses up to 4 more as references, 5 in total.
• An edit without aspect_ratio keeps the shape of the source image.
• resolution is 1K or 2K, and does not change the list price.

A budget rule

Treat the 7x gap as a budget rule. A 50-image test at low is $2.00, the same test at high is $14.00. Run the first pass at low, pick prompts that work, then spend high on the keepers. If you generate 200 finished images, the difference between all-medium ($16.00) and all-high ($56.00) is $40.00, which pays for a lot of low-tier exploration.

Compared with ChatGPT Image 2.5

ChatGPT Image 2.5 on Sume is 1 cent at low and 6 to 7 cents at high. Ideogram 4.5 low (4 cents) costs four times that low tier, and its high tier (28 cents) costs about four to five times that high tier. Price alone does not tell you which fits a given job, so run the same brief on both. Pin the quality in every request through the Image API and keep the model id explicit.

Check it before you run it

Every figure above is a catalog list price times 1.25, rounded up to the cent, as of 2026-10-08. Catalogs change, so before a large batch, read the current model entry in the docs and recompute the one line that matters for your case. Write the arithmetic next to the job in your own notes: list rate, seconds or characters, multiplier, rounding. If the result differs from the wallet charge by more than a cent, the catalog entry has changed, and the docs page is the place to find out why.

Run one small job first. Submit a single request with the pinned model id and the settings in the tables, poll the returned polling_url until it finishes, and compare the charge with your estimate. Then scale up. Using a pinned id for the test matters, because sume/auto never names the family, so you cannot tie its charge to the row you priced.