APIImage Generation
Seedream 4.5
Configurable fields and supported values for seedream-v4.5.
Model ID
seedream-v4.5
Capabilities
| Capability | Supported value |
|---|---|
| Text-to-image | Yes |
| Image-to-image | Yes |
| Multi-image reference | No |
Aspect ratio (aspect_ratio) | 1:1, 4:3, 3:4, 16:9, 9:16, 21:9 |
resolution | Not supported; do not send |
quality | low, medium, high |
| Reference field | image |
Supports one reference image through image.
Configurable fields
| Field | Required | Notes |
|---|---|---|
model | Yes | Must be seedream-v4.5. |
prompt | Yes | Text description or editing instruction. |
aspect_ratio | No | Preferred aspect-ratio field; select one supported value listed above. |
size | No | Compatibility alias for aspect_ratio. New integrations should use aspect_ratio. |
quality | No | Only use the quality values listed above. |
n | No | Defaults to 1; account/model limit comes from the model list API. |
image | No | Single reference image URL for image-to-image generation. |
response_format | No | url (default) or b64_json. |
output_format | No | png or jpeg. |
extra_body.optimize_prompt_options.mode | No | Set to standard to enable the provider's prompt optimization mode. |
wait_for_completion | No | Defaults to false, which returns an asynchronous task ID. Set to true to wait in the create request. |
polling_timeout | No | Polling timeout used only when wait_for_completion is true. |
Seedream prompt enhancement uses extra_body.optimize_prompt_options.mode. Do not send think_level, thinking_level, or reasoning_effort.
Example
{
"model": "seedream-v4.5",
"prompt": "Create a cinematic editorial image",
"image": "https://example.com/reference.png",
"aspect_ratio": "3:4",
"quality": "high",
"extra_body": {
"optimize_prompt_options": {
"mode": "standard"
}
}
}Last updated on