Music Generation

MLX-Audio music models generate complete songs from a musical caption and structured lyrics, or music and sound effects from a text prompt alone.

Model	Output	Models	Key features
MiniMax Music 3	44.1 kHz stereo	BF16, 8-bit, 6-bit, 4-bit, MXFP8, MXFP4, NVFP4	Multilingual lyrics, structural tags, long-form generation
Stable Audio 3	44.1 kHz stereo	small-sfx, small-music	Text to sound effects or music, up to 120 seconds

python -m mlx_audio.music.generate \
  --model mlx-community/MiniMax-Music3-mxfp8 \
  --caption "Warm acoustic pop with intimate female vocals" \
  --lyrics $'[verse]\nMorning light\n[chorus]\nSing with me' \
  --output song.wav