BytePlus ModelArk (direct)
# Seedream 4 ## Overview Seedream 4.0 called directly on BytePlus ModelArk. Text-to-image and image editing with multi-reference support. ## Model Metadata - Provider: byteplus - Category: static - Output Type: image - Slug: seedream-4-byteplus ## Notes Use the Playground tab above to test prompts and parameters.
Model input reference derived from preset schema.
| Name | Type | Required | Default | Description |
|---|---|---|---|---|
| size | select | No | "2K" | Image resolution: 1K (1024px), 2K (2048px), or 4K (4096px). |
| prompt | string | Yes | "" | Text prompt for image generation |
| max_images | integer | No | 1 | Max images to generate when sequential_image_generation='auto'. Input + generated cannot exceed 15. |
| image_input | image | No | [] | Input image(s) for image-to-image generation. Up to 14 reference images. |
| aspect_ratio | select | No | "match_input_image" | Aspect ratio hint (expressed via size/prompt on BytePlus; not sent as a separate field). |
| output_format | select | No | "jpeg" | seedream-4-0 output is jpeg only (custom output_format not supported). |
| sequential_image_generation | select | No | "disabled" | 'disabled' generates a single image. 'auto' lets the model generate a related set (input + generated <= 15). |