Qwen Image Max API on Sume: text-only, and how it differs
qwen/qwen-image-max on Sume is text-to-image only and rejects reference images with a 400. Price, ratios, a working call, and how it differs from qwen-image.

Call Qwen Image Max on Sume with model: "qwen/qwen-image-max". It is text-to-image only: a request with input_references returns 400 unsupported_parameter. One image is estimated at $0.10, against $0.03 for qwen/qwen-image, which does take references.
Read from the Image API docs and the catalog code on 2026-09-29. The docs do not describe how the two Qwen ids differ in output quality, so this post does not rank them.
How do the two Qwen ids differ on Sume?
The difference is in the request surface and the price, both readable from the catalog.
| Item | `qwen/qwen-image` | `qwen/qwen-image-max` |
|---|---|---|
| Reference images | Up to 10 | None; text-to-image only |
| Images per request | 1 to 4 | 1 to 4 |
| Estimated price per image | $0.03 | $0.10 |
output_format | png, jpeg, webp | png, jpeg, webp |
What does a working call look like?
aspect_ratio takes one of 13 listed ratios, including 1:1 and 21:9; custom pixels go in image_size. resolution and quality are not listed for this model.
curl -X POST "https://api.sume.com/v1/images" \
-H "Authorization: Bearer $SUME_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "qwen/qwen-image-max",
"prompt": "A hand-lettered chalkboard menu, blue and white",
"aspect_ratio": "3:4",
"n": 2
}'What happens if I add a reference image?
The request fails before generation with 400 unsupported_parameter, and the message says the model is text-to-image only. Nothing is billed. For edits, switch to qwen/qwen-image, which lists up to 10 references.
Does Qwen Image Max stream or return progress?
No image model streams in v1: stream: true returns 400 streaming_not_supported. Use mode: "async" and read the job events, or take a webhook. A synchronous call waits up to 30 seconds and returns 202 with a job if it is still running.
How is Qwen Image Max billed for several images?
Per image. The endpoint record has one output_image pricing line with Sume's margin already applied, so n images cost cost_usd times n. The call above with n: 2 is estimated at $0.20. The estimate is a list price rounded up to a cent.
A completed generation is billed in full; a failed or cancelled one is not. n is capped at 4 for this model, and a value outside 1 to 4 returns 400 invalid_request.
Which aspect ratios and sizes does Qwen Image Max take?
The catalog lists 1:1, 16:9, 9:16, 4:3, 3:4, 4:5, 5:4, 3:2, 2:3, 21:9, 9:21, 1:2, 2:1. A ratio outside that list returns 400 invalid_request with the accepted values in details.supported. Custom pixels go in image_size as a WIDTHxHEIGHT string or a { "width", "height" } object, since Sume treats Qwen as a custom-pixel family. The docs publish no pixel limits for it, so a size the provider refuses comes back as a failed generation, which is not billed.
Sources
Related posts
More in Models
- Recraft V4.1 API on Sume: check the catalog for the id
Sume lists recraft/recraft-v4, not a V4.1 id. How to check the live catalog for a newer Recraft, and what the listed model takes: text only, webp, $0.05.
- Recraft V4 SVG: can you get vector output through Sume?
Recraft says V4 can make editable SVG. Sume's recraft/recraft-v4 advertises webp output only. What that means, plus a working call.
- Remove shadow from photo with AI: edit it or cut it out
To remove a shadow from a photo with AI, send the photo to an image-edit model, name the one shadow to remove, and list what stays. Prompts, cost, limits.
- Runway Media Router and sume/auto: what Sume's router does
Looking for a Runway Media Router equivalent? Sume's video router is sume/auto: it picks the model, never says which ran, and clips run 3 to 10 seconds.
Written by Sume