GPT Image 2.5 background: transparent vs opaque in the API
Sume accepts background auto, transparent or opaque on the two GPT Image 2.5 ids only. What opaque does, what a bad value returns, and which models reject it.

Send background: "opaque" when you want GPT Image 2.5 to return an image with a solid background, and "transparent" when you want a cutout. "auto" leaves the choice to the model. Sume advertises background on openai/gpt-image-2.5 and openai/gpt-image-2.5-sunburst only, so any other image model returns a 400 if you send it.
Everything here is read from Sume's Image API docs and the catalog code on 2026-09-29. The docs list the three values but do not describe how each one renders, so this post does not either.
Which models accept the background field?
The field is part of the catalog descriptors of the two 2.5 ids. Every other image model leaves it out, and a request that sets a parameter the model does not list is rejected with 400 unsupported_parameter, not dropped.
| Model id | Accepts `background` | Values |
|---|---|---|
openai/gpt-image-2.5 | Yes | auto, transparent, opaque |
openai/gpt-image-2.5-sunburst | Yes | auto, transparent, opaque |
openai/gpt-image-2 | No | 400 unsupported_parameter |
| Other catalog models | No | 400 unsupported_parameter |
What does an opaque request look like?
Add background next to the fields you already send. It is independent of output_format and quality, which are separate fields in the request table.
curl -X POST "https://api.sume.com/v1/images" \
-H "Authorization: Bearer $SUME_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "openai/gpt-image-2.5",
"prompt": "a single ceramic mug on a plain studio backdrop",
"background": "opaque",
"output_format": "png"
}'What happens if I send another value, like white?
Sume checks background against the model's listed values. Anything outside auto, transparent and opaque returns 400 invalid_request, and details.supported lists the accepted values. To get a white backdrop, describe it in the prompt and send opaque; there is no color field.
Should I send auto, opaque or transparent?
Send opaque or transparent when your pipeline depends on the result, such as a product tile that must sit on your own color, or a sticker that must have no backdrop. Leave auto for exploration. For the transparent side, including the output-format catch, see GPT Image 2.5 transparent background.
If a request fails, the job or response error envelope is described in Errors and rate limits.
How do I confirm which background I got?
Check the response, not just the picture. Images come back as Sume-hosted URLs, and each entry carries a media_type such as image/png or image/webp, matching the output_format you asked for. Open the file on a dark and a light backdrop: a solid fill and a cutout look the same on a white page.
usage.cost is the USD amount billed to your wallet. The docs tie their 2.5 price notes to output size and quality and do not mention background, so this post does not assign it a price. Billing is all-or-nothing: a completed generation is billed in full and a failed one is not.
Sources
Related posts
More in Developers
- GPT Image 2.5 SVG output: why the API rejects svg
GPT Image 2.5 on Sume advertises png, jpeg and webp. Asking for svg returns a 400 that lists the accepted formats. How to check and what to do for vector art.
- GPT Image 2 vs GPT Image 2.5: what changes when you migrate
GPT Image 2 stays selectable on Sume. Moving to 2.5 means a new model id and new limits: six quality values, 16 references, mask_url. Old requests keep working.
- GPT Image 2.5 4K: how to request a 3840x2160 image by API
To get a 4K image from GPT Image 2.5 on Sume, send image_size 3840x2160 to POST /v1/images and be ready for a 202 job response. Request, cost, and polling.
- GPT Image 2.5 image editing API: edit a photo with a prompt
Edit a photo with GPT Image 2.5 on Sume: send the image in input_references, describe the change, and set aspect_ratio to auto. Up to 16 references per call.
Written by Sume