bzikst/minimax-music3

MiniMax-Music3: full songs from lyrics and a style description. Serves the official sglang-omni pipeline and returns stereo m4a/AAC by default. Detects available GPUs and colocates or splits the autoregressive and acoustic stages accordingly.

Public
197 runs

Run bzikst/minimax-music3 with an API

Use one of our client libraries to get started quickly. Clicking on a library will take you to the Playground tab where you can tweak different inputs, see the results, and copy the corresponding code to use in your own project.

Input schema

The fields you can use to run this model with an API. If you don't give a value for a field its default value will be used.

Field Type Default value Description
lyrics
string
Song lyrics. Section tags ([verse], [chorus] and others) must each be on their own line.
caption
string
Complete style description. When set, the three fields below are ignored.
global_metadata
string
Tempo, key, genre, emotional arc, overall sonic character.
vocal_details
string
Vocal gender and timbre, delivery style, backing vocals, processing.
arrangement
string
Instruments per section, rhythmic development, textures and transitions.
duration_seconds
integer
300

Min: 5

Max: 360

Upper bound on duration. The model ends the song on its own, almost always before the bound.
seed
integer
Generation seed. Empty means random. The same seed with the same input yields a byte-identical track.
response_format
None
m4a
Output format.

Output schema

The shape of the response you’ll get when you run this model with an API.

Schema
{
  "type": "string",
  "title": "Output",
  "format": "uri"
}