AI Research Atlas

Lyria 3 in the Gemini app

Google · 18 February 2026

Lyria 3 brings text- or image-prompted 30-second music tracks with vocals and cover art to the Gemini app in beta.

Google DeepMind's most advanced music model reaches consumers. All outputs carry SynthID watermarks. Lyria 3 Pro (2026-03-25) extends tracks to three minutes with structural control; Lyria 3.5 followed in July.

Date
Wednesday, 18 February 2026
Lab
Google
Kind
model
Access
app only

Sources

  1. deepmind.google/blog/a-new-way-to-express-yourself-gemini-can-now-create-music/
  2. deepmind.google/blog/lyria-3-pro-create-longer-tracks-in-more/
  3. blog.google/intl/en-ie/products/lyria-3/
  4. ai.google.dev/gemini-api/docs/changelog

This record was checked against its sources on 6 October 2026. How we check

Related