Google DeepMind Launches Lyria 3: AI Music Generation Now Available in Gemini

Google DeepMind is rolling out custom music generation to the Gemini app through Lyria 3, its latest generative music model. The feature, now in beta, lets users create original 30-second tracks by describing an idea or uploading a photo.

How It Works#

The system supports two primary workflows:

Text to track: Describe a genre, mood, or memory—such as "a comical R&B slow jam about a sock finding their match"—and Lyria 3 generates a full track with original lyrics and instrumentation in seconds.

Photo and video to track: Upload visual content and let Gemini compose a track that matches the mood. For example, users can request music about a pet on a hiking adventure.

Each track includes AI-generated cover art via Nano Banana, making it shareable via download or link.

Key Improvements#

Lyria 3 builds on earlier Lyria models with three core upgrades:

• Auto-generated lyrics: No need to write lyrics yourself; they're created based on your prompt
• Creative control: Customize style, vocals, tempo, and other musical elements
• Musical complexity: More realistic and intricate track composition

Responsible AI & Content Verification#

Google has embedded SynthID—an imperceptible watermark—in all Gemini-generated tracks to identify AI-created audio. The platform also expanded its verification tools to let users check whether audio, images, or video were created by Google AI. "We've sought to develop this technology responsibly in collaboration with the music community," DeepMind notes, emphasizing that music generation is designed for original expression, not artist impersonation.

If a prompt names a specific artist, Lyria 3 treats it as creative inspiration rather than mimicry, generating tracks with similar style or mood instead. The system also filters outputs against existing content, and users can report potential copyright violations.

Availability & Expansion#

The feature rolls out today for users 18+ in English, German, Spanish, French, Hindi, Japanese, Korean, and Portuguese. It's launching on desktop immediately and coming to mobile "over the next several days." Google AI Plus, Pro, and Ultra subscribers get higher generation limits.