Seedance, Kling, MiniMax H3 and Wan video models
Choose ten video models through BetterToken for text, image and reference generation plus H3 regeneration, with one asynchronous API.
2026-11-06
Use one BetterToken API Key for Seedance, Kling, MiniMax H3 and Wan, with shared quoting, submission, polling and download endpoints. Choose a model for your text, images or reference media.
Models and operations
| Model | Model ID | Supported operations |
|---|---|---|
| Seedance 2.5 CN | seedance-2.5-cn | text_to_video, image_to_video, reference_to_video |
| Seedance 2.0 CN | seedance-2.0-cn | text_to_video, image_to_video, reference_to_video |
| Seedance 2.0 Fast CN | seedance-2.0-fast-cn | text_to_video, image_to_video, reference_to_video |
| Seedance 2.0 Mini CN | seedance-2.0-mini-cn | text_to_video, image_to_video, reference_to_video |
| Kling 3.0 | kling-3.0 | text_to_video, image_to_video |
| MiniMax H3 | minimax-h3 | text_to_video, image_to_video, reference_to_video, video_edit |
| Wan 3.0 | wan-3.0 | text_to_video, image_to_video, reference_to_video |
| Wan 3.0 Prime | wan-3.0-prime | text_to_video, image_to_video, reference_to_video |
| Wan 2.7 T2V | wan-2.7-t2v | text_to_video |
| Wan 2.7 I2V | wan-2.7-i2v | image_to_video |
All four Seedance models in this batch are China variants. Keep -cn in the Model ID; do not replace it with a different unsuffixed version.
Choose by input
text_to_video: Generate video from text.image_to_video: Generate from a first frame or first/last frames.reference_to_video: Generate with image, video or audio references; supported media depend on the model.video_edit: Regenerate an eligible existing 768p H3 task at 2K.
Get started
Integration, quote, polling and download examples
Quotes and charges
Charges depend on the model, duration, output tier, audio and reference media. USD and RUB prices are independent. Reserved amounts and maximum charges in quotes are not the final amount charged; read the task’s settlement information.
Current models and pricing
