Minimax Music 3

Generate songs up to 5 minutes long, vocals included.

MiniMax Music 3: Full Songs from Lyrics and a Description

MiniMax Music 3 generates complete songs up to 5 minutes long, vocals included. You give it two things: a description of the music (genre, mood, instrumentation) and the lyrics you want sung. It writes the melody, arranges the instruments, and performs the vocals in one pass.

The output is studio-format audio: 44.1 kHz 16-bit stereo WAV, ready to master, mix, or drop straight into a video.

What can you do with it?

How to use it

  1. Describe the music. In the Music tab, write what the song sounds like. The model responds best to structured captions: genre, BPM, key, the emotional arc, what the vocals sound like, and how the sections build.
  2. Write the lyrics. In the Lyrics tab, use structure tags on their own lines: [intro], [verse], [pre-chorus], [chorus], [post-chorus], [bridge], [instrumental], [solo], [outro].
  3. Set the duration. 1 to 300 seconds, default 60. It’s an upper bound; the song ends when the arrangement does.
  4. Run it. One track comes back as stereo WAV.

What you get

One complete song as a 44.1 kHz 16-bit stereo WAV file, up to 5 minutes. Vocals follow your lyrics line by line; structure follows your tags.

Tips

All AI models & tools

See pricing and start creating on Mitte