Lyria is Google DeepMind's family of music generation models, which create songs and instrumental tracks from text descriptions or images.
In plain words
Music is hard for AI because it must stay coherent across beats, phrases and whole verses. Google DeepMind announced Lyria in November 2023 with YouTube. The current version, Lyria 3.5, improves vocals and lyrics and can make tracks of varying length. Through the Gemini API it produces high-quality stereo audio from a text prompt or an image, and a separate model, Lyria RealTime, streams music live as you steer it.
A quick example
Ask the clip model for a gentle solo guitar tune with no vocals. It returns a 30-second audio clip you can play or save; the full Lyria 3.5 model can make longer tracks.
Where you'll meet it
- The Gemini API and Google's music tools.
- Google Flow Music.
Related
Sources
- DocsTransforming the future of music creation, Google DeepMind · read 29 Sept 2026
- DocsGenerate music with Lyria 3.5, Gemini API docs · read 2 Nov 2026
- DocsLyria, Google DeepMind · read 29 Sept 2026
Checked by WhatIsAI against the sources above · 2 Nov 2026. Spotted an error? Report it.