Seedream 5.0 Pro API: Model IDs, Endpoints, and Code Samples

Seedream 5.0 Pro is ByteDance's flagship image generation and editing model, released in July 2026. This guide covers every way to call it from your own code, with the exact model IDs and request formats.

The fastest way: generate here first

Before wiring up an API, you can try Seedream 5.0 Pro directly in our generator — same model, no setup. Free credits are included when you register.

Official platforms

Volcengine Ark (China)

Model ID: doubao-seedream-5-0-pro-260628
API reference: Volcengine docs
Tutorial: Volcengine tutorial

BytePlus ModelArk (International)

Model ID: dola-seedream-5-0-pro-260628
API reference: BytePlus docs
User guide: BytePlus guide

Both expose the model through the images generations API. Text-to-image and image editing share one endpoint — pass reference images to switch into editing mode.

Aggregator access (no ByteDance account)

Aggregators like Kie.ai resell the model per-image with no enterprise onboarding:

Text to image: seedream/5-pro-text-to-image
Image to image / editing: seedream/5-pro-image-to-image

curl -X POST https://api.kie.ai/api/v1/jobs/createTask \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "seedream/5-pro-text-to-image",
    "input": {
      "prompt": "A product poster with the headline \"Summer Drop\" in crisp small type, minimal studio lighting",
      "aspect_ratio": "3:4",
      "quality": "high"
    }
  }'

The task is asynchronous: the response returns a taskId which you poll (or receive via callBackUrl webhook) until the image URL is ready.

Parameters that matter

Parameter	Values	Notes
`aspect_ratio`	`1:1`, `4:3`, `3:4`, `16:9`, `9:16`, `2:3`, `3:2`	—
`quality`	`basic` (1K), `high` (2K)	16:9 and 9:16 do not support 2K
`image_urls`	up to 10 URLs	JPEG/PNG/WebP, ≤10 MB each; switches to editing mode
`prompt`	3–3000 chars	supports native input in 14 languages

One request produces one image — Seedream 5.0 Pro does not support sequential or streaming generation.

What's new in 5.0 Pro

Interactive precision editing — point, box, and anchor selection; edits land exactly where you indicate and the rest of the image stays identical
Layer separation — split an image into background plus element layers, each a transparent-alpha PNG
Sharper small text — usable at 1K, noticeably better at 2K; slides, posters, and infographics become practical
14-language native text — English, Chinese, Russian, Japanese, Korean, German, French, Arabic, Spanish, and more, with correct letterforms