MiniMax Music FAQ

MiniMax Music is a family of AI music generation models. It can turn text prompts and lyrics into vocal songs or instrumental tracks. You can describe the genre, mood, instruments, scene, and vocal direction before generating.

Choose MiniMax Music 3.0 when you want detailed direction for instruments, playing style, vocals, and arrangement. Music 2.6, 2.5, and 1.5 remain available as earlier workflows. Choose Music Cover when you want to create a new version from reference audio.

Describe the sound you want, add lyrics for a vocal song, and choose your audio settings. You can also turn on Instrumental Track to create music without singing. Submit the task, wait for it to finish, then preview and download the result.

For a vocal song, lyrics are required and the music prompt is optional. For an instrumental track, the prompt is required and lyrics are not needed. Music 3.0 supports prompts up to 2,000 characters and vocal-song lyrics up to 3,500 characters.

Supported tags include [Intro], [Verse], [Pre Chorus], [Chorus], [Interlude], [Bridge], [Outro], [Post Chorus], [Transition], [Break], [Hook], [Build Up], [Inst], and [Solo]. Put each tag on its own line to organize the song.

Yes. Turn on Instrumental Track and describe the style, mood, scene, instruments, and performance you want. Lyrics are not required in instrumental mode.

Music 3.0 supports MP3, WAV, and PCM output. Available sample rates are 16, 24, 32, and 44.1 kHz. Available bitrates are 32, 64, 128, and 256 kbps.

Generation usually takes about five minutes, but the exact time can vary with service demand and task complexity. Keep the page open while the task is processing.

Each MiniMax Music 3.0 generation currently costs 5 credits on this site. The button shows the credit cost before you submit the task.

Yes. When the task is complete, you can preview the track in your browser and use the Download Music button to save the audio file to your device.

Yes. You can use generated music commercially. Review the applicable terms and each generated track before publishing or distributing it.

No. Music Cover is a separate workflow for transforming reference audio into a different target style. Music 3.0 creates new songs or instrumentals from prompts and lyrics.