Google Rolls Out Lyria 3 in Gemini App for Instant AI Music Tracks
Google DeepMind launched Lyria 3 on February 18, 2026, integrating the model into the Gemini app to enable users to generate 30-second music tracks from text prompts or images.
 |
| Credits: Google |
The rollout makes the technology available globally to individuals aged 18 and older who access the Gemini app through personal Google accounts or Workspace subscriptions.
Users in English, German, Spanish, French, Hindi, Japanese, Korean, and Portuguese can input descriptions such as an upbeat birthday tune or an Afrobeat track recalling family memories, prompting the system to produce complete audio with instrumentals, vocals, and automatically generated lyrics.
Image uploads allow the model to interpret visual elements and convert them into corresponding soundtracks, while video clips serve as references for matching vibes in the output.
The generation process handles various genres including pop, funk, and Motown, with options to specify tempo, dynamics, vocal styles, and acoustic preferences for customized results.
Google DeepMind developed Lyria 3 with feedback from producers and musicians to capture elements of musicality from rhythm to arrangement, allowing the creation of cohesive tracks that flow naturally from start to finish.
"Lyria 3 is our most advanced music generation model yet," states the official model page.
Free access provides basic usage, but Google AI Plus, Pro, and Ultra subscribers receive expanded limits for more frequent generations.
Safety measures incorporate extensive content filtering to reduce harmful outputs, privacy protections in audio processing, and SynthID watermarking to identify AI-generated music through imperceptible embeds.
The model supports exporting professional-grade audio suitable for projects like background scores or presentations, though users must verify tracks align with their intended vision due to ongoing improvements in capabilities.
Integration extends to YouTube Shorts under the Dream Track feature for AI soundtracks, broadening applications across Google platforms.
