kinoAPIUse the right model for the next frame. These are KinoAPI integration parameters, not an exhaustive list of the provider’s capabilities.
A new chapter in video.
ByteDance describes Seedance 2.5 as a generation with 30-second storytelling, more precise references and editing. KinoAPI currently exposes 480p and 720p, with text/image input and a dedicated video-reference input route.
Provider model overview ↗| Model / ID | Output tiers | Duration / type | Use case |
|---|---|---|---|
Seedance 2.5seedance-2.5 | 480p / 720p | 4–30s / Auto | Text and image input. Token-based billing. |
Seedance 2.5 · Video inputseedance-2.5-video | 480p / 720p | Auto / adaptive | Video references; a separate price class, not a different underlying model. |
Seedance 2.0 Proseedance-2.0-pro | 480p–4K | 4–15s | Existing high-resolution video workflow. |
Seedance 2.0 Fastseedance-2.0-fast | 480p / 720p | 4–15s | Existing fast video workflow. |
Seedance 2.0 Miniseedance-2.0-mini | 480p / 720p | 4–15s | Drafts and iterations. |
Seedance 1.5 Proseedance-1.5-pro | 480p / 720p | 4–15s | Legacy model with an optional audio price tier. |
Seedream 5.0 Proseedream-5.0-pro | 1K / 2K | Image | Detailed image generation, synchronous response. |
Seedream 5.0 Liteseedream-5.0-lite | 2K / 3K / 4K | Image | Lower-cost image creation and iteration. |
Configuration checked September 10, 2026. Call GET /v1/models before generation. A listed parameter range is not a guarantee that every combination has been independently tested.
KinoAPI is an independent relay for video and image generation, with a private cloud canvas. Use Seedance through an asynchronous video API and Seedream through a synchronous image API, with one KinoAPI account and balance. It is not affiliated with ByteDance or Volcano Engine.
Seedance 2.5 is billed by actual reported tokens: $15.61875 per million tokens without video input, or $9.345 per million tokens with video input. Video-input usage can include the reference. These are not per-second prices. See the pricing page and your final task record.
Use seedance-2.5 for text or image input. Use seedance-2.5-video when metadata.content contains a video reference; this mode currently requires adaptive ratio and Auto duration (-1). Both route to Seedance 2.5. Check GET /v1/models for availability on your key.