Ideogram V3 vs 4.5 on Sume: both $0.075, 10 vs 5 references

Answer

Ideogram V3 and Ideogram 4.5 are both listed at $0.075 per image on Sume, but they differ in references and quality. V3 takes up to 10 reference images and maps quality to the provider's rendering speed. Ideogram 4.5 takes up to 5 references, adds a resolution of 1K or 2K, and prices by quality tier: $0.0375 low, $0.075 medium and $0.275 high.

So the $0.075 match only holds at medium on 4.5. If you set quality: high on 4.5 the same prompt costs $0.275, which is 3.7 times the V3 price.

How editing differs

On 4.5, a request with input_references edits the first image and treats up to four more as references. An edit without aspect_ratio keeps the shape of the source image. On V3 the catalog lists up to ten reference images, so a brand kit with a logo, a palette card and several product shots fits in one call.

The Ideogram 4.5 tier prices are from the fal Ideogram v4.5 page (read 2026-10-04): $0.03, $0.06 and $0.22 per image by quality, which Sume multiplies by 1.25 to give $0.0375, $0.075 and $0.275.

Choosing

• Many references or a loose brand board: V3.
• A cheap text-in-image draft: 4.5 at low, then rerun winners at medium or high.
• Print or large crops: 4.5 at resolution: 2K.
• Not sure: send sume/auto and let the image router choose, but then you cannot pin the family.

Both rows list the same 15 ratios, including 1:3 and 3:1. Neither accepts mask_url or background; those belong to ChatGPT Image 2.5. See the Sume Image API docs.

Before a large run

Prices and descriptors change when the catalog changes, so confirm them before you spend. Call GET /v1/images/models/{id}/endpoints for the row you plan to use and read its pricing line and supported_parameters; both come back in one response.

Then run a pilot of three to five images and read usage.cost on each response. Multiply by your planned count for a forecast you can trust. Completed generations are billed in full and failed or cancelled ones are not, so a pilot that errors costs nothing.

For big batches, use mode: "async" or mode: "webhook" with a public HTTPS webhook_url, so no request waits on the 30-second sync limit. Poll GET /v1/jobs/{id}/status and fetch GET /v1/jobs/{id}/result when the job completes.