Models
Seedance 2.5
Model Overview
Direct supply
Video GenByteDance Seedance 2.5 — professional multimodal video generation with text, image and video references; 480P-1080P, billed per million output video tokens.
Input Modalities
Text
Output Modalities
Video
Recommended for
Suited for creative workflows that generate video from input.
Capabilities
Video generation
Generates video from a text prompt or an image
Supported Modes
Video Generation
Model Pricing
10% off
Pay-as-you-go
Priced per spec
Seedance 2.5 is priced per spec — each tier bills at its own rate, resolved from the request.
480P · No video input$10.42$9.378per 1M tokens
480P · With video input$6.252$5.627per 1M tokens
720P · No video input$10.42$9.378per 1M tokens
720P · With video input$6.252$5.627per 1M tokens
1080P · No video input$11.462$10.316per 1M tokens
1080P · With video input$6.848$6.163per 1M tokens
Billing notes
- Reference video is accepted and not billed.
- Reference images are billed per image.
- Reference audio is not supported.
- Failed tasks are not billed.
Indicative pricing — final cost is confirmed at request time.
API Code Example
HTTP
import requests
resp = requests.post(
"https://api.modelsite.ai/api/v3/contents/generations/tasks",
headers={"Authorization": "Bearer YOUR_API_KEY"},
json={
"content": [
{
"text": "A cat playing the piano",
"type": "text"
}
],
"model": "dreamina-seedance-2-5"
},
)
print(resp.json())
# The task id is at id in the response; poll it until the video is ready.
task_id = resp.json()["id"]
poll = requests.get(
f"https://api.modelsite.ai/api/v3/contents/generations/tasks/{task_id}",
headers={"Authorization": "Bearer YOUR_API_KEY"},
)
print(poll.json())