Seedance 2.5
A multimodal video generation model supporting text, image, video, and audio references
About Seedance 2.5
Seedance 2.5 is ByteDance's video generation model, providing text-to-video, image-to-video, start-and-end frame control, multimodal references, and prompt-driven video editing through the reAPI asynchronous API. It generates 4–30-second videos per request, up to 30 seconds, supports 480p, 720p, and 1080p, and can produce synchronized speech, sound effects, and music. Each request accepts up to 30 images, 10 videos, and 10 audio tracks, and also supports audio-only input.
Best for
Best use cases
Text-to-video generation
Generation from image, video, or audio references
Start-and-end frame control
Prompt-driven video editing
Features
Multimodal references
Each request accepts up to 30 images, 10 videos, and 10 audio tracks, and also supports audio-only input.
Five generation modes
Supports text-to-video, image-to-video, start-and-end frames, multimodal references, and prompt-driven video editing.
Long-form video output
Generates videos of 4–30 seconds per request, with a maximum length of 30 seconds.
Synchronized audio
Output can include synchronized speech, sound effects, and music, with support for 480p, 720p, or 1080p.
Pros
- Supports multimodal image, video, and audio references
- Supports 480p, 720p, and 1080p with synchronized audio
- Failed tasks are not charged, and reserved credits are automatically returned
Cons
- You are responsible for the rights and consent requirements when uploading reference assets containing real people
Pricing Plans
Usage-based billing
- Billed based on completed generations
* Pricing information is for reference only. Please refer to the latest announcement on Seedance 2.5's official website.