Seedream 5.0 pro layer decomposition vs a Sume Seedream result
Seedream 5.0 pro can split one image into a base plus up to 16 PNG layers. Sume lists Seedream 5.0 Lite, and results are one Sume-hosted URL per image.

Layer decomposition is a Seedream 5.0 pro feature: with layer_decomposition set to true and exactly one input image, BytePlus returns a base image plus up to 16 independently editable PNG layers, each with a z_index. Sume's catalog lists Seedream 5.0 Lite, not Pro, and returns one Sume-hosted URL per image, so this is not available there.
BytePlus facts are from its Seedream 5.0 overview; Sume facts from the Image API docs and catalog code, read 2026-09-30.
What does layer decomposition return?
The overview says each layer is a PNG with an alpha channel, and the response includes each layer's position, stacking order (z_index) and a content description. The prompt is optional and can name elements to separate. It reserves 17 IPM per request (one base plus up to 16 layers) and returns unused quota afterwards. The overview lists it as not available on 5.0 lite or the 4.x models.
What does a Sume Seedream request return?
The catalog entry is seedream-5-lite (Seedream 5.0 Lite), with Seedream 4.5 alongside it. Images return as Sume-hosted URLs, data[].url and signed, rather than inline base64, with usage data. There is no layers array in the documented payload. If you send a parameter the model does not list, you get 400 unsupported_parameter, so do not send layer_decomposition.
| Item | Seedream 5.0 pro (BytePlus) | Sume Seedream ids |
|---|---|---|
| Layers | Base plus up to 16 PNG layers | Not documented |
| Input for the mode | Exactly one image | n/a |
| Output shape | Layers with z_index | data[].url, one per image |
| Listed variant | Pro | 5.0 Lite (seedream-5-lite) and 4.5 |
How do I get separate elements on Sume?
Generate elements as separate requests, or edit one region at a time. Sume documents no alpha-channel layer output for Seedream, so any compositing happens in your own tool. Check the catalog before choosing: Seedream 5.0 Pro vs Lite on Sume covers the ids.
Sources
Related posts
More in Models
- Suno v6 audio and video input vs Sume: text and one image
Suno v6 says you can create with text, audio, images and video. Sume's music request takes a text prompt and one optional image; audio and video are not inputs.
- Suno v6-wild: is there a less predictable setting on Sume?
Suno v6-wild is a paid model that is less predictable. Sume's music request has no seed, temperature or guidance field, so variety comes from the prompt.
- Wan 3.0 on Runway, or the wan-3.0 id as an API job on Sume
Runway's changelog lists Wan 3.0 in tool mode and workflows on paid plans. On Sume, wan-3.0 is a model id you call from code and poll as a job.
- An OpenRouter-compatible video API: sume/auto or a pinned model
Sume's POST /v1/videos follows OpenRouter's video generation API field for field. Let sume/auto pick the model, or pin a catalog id like seedance-2.5.
Written by Sume