Models
List the video, image and music models available through the Seedance 3.0 API, with their scenes, durations, resolutions and credit cost.
Use GET /models to get the live catalog. Pass a model's id as model when you create a generation.
curl https://seedance3ai.video/api/v1/modelsNo API key is required for this endpoint.
Response
{
"object": "list",
"data": [
{
"id": "seedance25-text-to-video",
"object": "model",
"name": "Seedance 2.5",
"media_type": "video",
"scenes": ["text-to-video"],
"description": "…",
"duration": "4-15 sec",
"credits": "…",
"has_audio": true,
"parameters": {
"durations": ["4", "5", "6", "…", "15"],
"default_duration": "5",
"resolutions": ["480p", "720p", "1080p"],
"default_resolution": "720p",
"aspect_ratios": ["16:9", "9:16", "1:1", "4:3", "3:4", "21:9"],
"default_aspect_ratio": "16:9",
"credits_per_second": { "480p": 49, "720p": 104, "1080p": 260 }
}
}
]
}| Field | Description |
|---|---|
id | Value to send as model |
media_type | video, image or music |
scenes | Input modes the model supports (see below). The first one is used when you omit scene |
duration | Human-readable output length range |
credits | Starting credit cost shown on the website |
has_audio | Whether the model can generate a soundtrack |
parameters | Present on catalog video models: the exact accepted duration, resolution, aspect_ratio |
Scenes
| Scene | Input |
|---|---|
text-to-video | Prompt only |
image-to-video | Prompt plus a first-frame image |
reference-to-video | Prompt plus reference images, videos or audio |
video-to-video | Edit or transform an existing video |
audio-to-video | Drive a video from an audio track |
text-to-image | Prompt only |
image-to-image | Prompt plus reference images |
text-to-music | Prompt only |
Most model families ship one id per scene, for example seedance25-text-to-video, seedance25-image-to-video and seedance25-reference-to-video.
Popular video models
| Model id prefix | Family | Notes |
|---|---|---|
seedance25- | Seedance 2.5 | 4–15 s, up to 1080p, native audio |
seedance25-flash- | Seedance 2.5 Flash | Faster and cheaper, up to 30 s at 720p |
seedance2mini- | Seedance 2.0 Mini | Lowest cost, up to 720p |
kling4- | Kling 4 | Strong motion and camera control |
wan27- | Wan 2.7 | 1080p, end-frame control |
happyhorse- | HappyHorse 1.0 | Text, image, reference and video editing |
ltx23- | LTX 2.3 | Up to 4K, audio-to-video, extend and retake |
The catalog changes as models are added or retired — always read ids from GET /models instead of hard-coding the list above.
Cost
Credits are charged when a generation starts and refunded automatically if it fails (except when the finished output is blocked by moderation, or the job times out at the provider). For catalog video models the cost is credits_per_second[resolution] × duration. Check your balance with GET /credits.