Generate music with Music 3.0
MiniMax Music 3.0 — the latest MiniMax music model: full-length songs from a style prompt and/or your own lyrics. Lyrics optimizer can write or polish lyrics from the prompt; instrumental mode skips vocals. WAV or MP3 output.
Asynchronous. Returns a job id (the track is not in this response). Poll GET /v1/audios/{id} until status is succeeded — the track URL is in that response’s audio_url.
Authorizations
API key as a bearer token: Authorization: Bearer bpx_live_xxx
Body
Output audio format. Defaults to wav (lossless); choose mp3 for a smaller file.
wav, mp3 Generate an instrumental track (no vocals).
Lyrics to sing (1-3500 characters). Required unless instrumental is true or lyrics_optimizer is true (the optimizer writes lyrics from the prompt).
Let the model write or polish lyrics from the prompt. Default true.
Style/mood description of the track (up to 2000 characters). Optional when lyrics are provided.