Skip to content

MusicLM

Generate high-fidelity music from a text description

Visit official site

What MusicLM is

The model is described as a text-to-audio system that generates high-fidelity music matching a text description.

What you can do with it

  • Generate high-fidelity music from text
  • Produce 24 kHz audio
  • Condition generation on a melody

Model notes

  • MusicCaps is named as the dataset
  • Audio quality and text consistency are described as improved over earlier systems
  • Example uses include music clips and music from a painting description

Pros & cons

✓ What we like

  • Audio quality and text consistency are described as above earlier systems
  • MusicCaps is named as the dataset

! What to watch out for

  • Generation is driven by a text description or a melody
  • The described output is audio at 24 kHz

FAQ

What sample rate is named?

The material states 24 kHz audio.

What dataset is used?

MusicCaps is named in the material.

Last reviewed: 2026-11-15

More AI music creation tools

View all →

How we review