Flux 2 Pro vs Flux 2 Flex on Sume: $0.0375 vs $0.0625

On Sume, Flux 2 Pro costs $0.0375 per image and Flux 2 Flex $0.0625. Both take 10 references, n up to 4 and 13 aspect ratios. 1,000 images: $37.50 vs $62.50.

4 min readSume
All posts

Answer

Flux 2 Pro costs $0.0375 per image on Sume and Flux 2 Flex costs $0.0625, so a thousand images cost $37.50 and $62.50. The two rows list identical parameter descriptors: up to 10 reference images, n from 1 to 4, and the same 13 aspect ratios, so the only measured difference in the catalog is the price.

Flex costs 67 percent more per image. If your prompts return the look you want on Pro, there is no catalog reason to pay for Flex.

Flux 2 rows in the Sume catalog and cost at volume (read 2026-10-04)
FieldFlux 2 ProFlux 2 Flex
Model idblack-forest-labs/flux.2-problack-forest-labs/flux.2-flex
Price per image$0.0375$0.0625
100 images$3.75$6.25
1,000 images$37.50$62.50
Reference imagesup to 10up to 10
n1 to 41 to 4
Aspect ratios13, 1:2 to 21:9 and 9:2113, same list
quality, resolutionnot offerednot offered

A fair test

Run the same 20 prompts on both rows with fixed settings, then compare the outputs side by side. Sume does not serve a seed field in v1, so results vary between runs and you should generate at least two takes of each prompt. Send n: 2 on each row, which makes 40 images per model: $1.50 on Pro and $2.50 on Flex.

Judge on what you actually ship: the crop, the text legibility, the product shape. A model that wins on a landscape prompt may lose on an object edit, so use your own brief rather than a generic one.

Request

curl -X POST https://api.sume.com/v1/images \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"black-forest-labs/flux.2-flex","prompt":"a linen sofa in a sunlit room","aspect_ratio":"3:2","n":2}'

Neither Flux row takes quality, resolution, mask_url or background; sending any of them returns a 400 unsupported_parameter. Reference for the full table: 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.

Sources

Related posts

More in Comparisons

All Comparisons posts

Written by Sume