Google has integrated its Lyria 3 AI model into the Gemini app, enabling users to create 30-second music tracks from simple prompts. The feature, which also generates lyrics and album art, is rolling out today with safeguards like watermarking to identify AI content. It expands Gemini's capabilities beyond text, images, and video.
Google announced on February 18, 2026, that its Gemini app now incorporates the Lyria 3 model, allowing users to generate short music clips. With a prompt like "a comical R&B slow jam about a sock finding their match," Gemini produces 30-second tracks, including automatically generated lyrics, without requiring users to provide them. The model also supports remixing existing tracks and creating music from photos or videos, while offering control over elements such as tempo and drumming style.
Lyria 3 builds on previous versions by producing more realistic and musically complex outputs. Each generation includes album art created by Google's Nano Banana image model. The feature extends to YouTube's Dream Track for generating backing tracks in Shorts, pairing with AI video tools like Veo.
Google provides example prompts and resulting tracks, such as "Sweet Like Plantain," a fun afrobeat track about childhood memories of home-cooked plantains, and "Sea Shanty," an a cappella piece with male choir vocals, foot-stomps, and handclaps simulating a ship's deck. Other examples include a Motown parody with orchestral R&B elements and a wistful pop track with breathy vocals and rainy city vibes.
To address concerns over AI music, tracks are watermarked with SynthID, an audio identifier detectable via Gemini, similar to tools for images and videos introduced at Google I/O 2025. Google states that naming a specific artist in prompts provides broad inspiration rather than direct imitation, though it acknowledges limitations and encourages reporting of issues. The model respects copyright and partner agreements.
Availability begins today in the Gemini web interface for users aged 18 and older, supporting English, Spanish, German, French, Hindi, Japanese, Korean, and Portuguese. Mobile app rollout follows in a few days, with higher usage limits for AI Pro and AI Ultra subscribers. More languages are planned soon.
