MiniMax API v1 updates:

  • POST llm now supports the MiniMax-M1 reasoning model, allowing up to 1M-token input and 80K-token output.
  • POST videos/create now supports the Hailuo 02 model which generates videos at 768p (6 or 10 seconds) and 1080p (6 seconds) resolutions, features advanced instruction following, handles complex physics (including acrobatics 🤹‍♀️), and supports native 1080p output.

Note:
The DeepSeek R1 model not longer supported.

Examples

fileID start frame

curl --location 'https://api.useapi.net/v1/minimax/videos/create' \
--header 'Authorization: Bearer user:12345-…' \
--form 'model="02"' \
--form 'options="768p-10sec"' \
--form 'prompt="The lady throws cupcake  at the camera. The cupcake hits the camera lens, completely blocking the view as it smashes. Then, as the cupcake slowly slides down the camera, we see the lady jumping with joy in the background."' \
--form 'fileID="user:12345-minimax:…-file:…"'

Result