Google has expanded its AI music generation ecosystem with the launch of Lyria 3 and Lyria 3 Pro, making the models available through the Gemini API, Google AI Studio, and several other Google platforms. The new models are designed to generate high-fidelity music with structured compositions, targeting developers, musicians, and content creators.
Lyria 3: Studio-Quality Music Generation
Lyria 3 is built to produce full musical compositions with elements such as vocals, verses, choruses, and bridges. The model supports multiple genres and languages, including styles like pop, funk, and Motown.
Google is offering two variants of the model:
Lyria 3 Pro (lyria-3-pro-preview)
- Generates full tracks up to three minutes long
- Supports prompts for song structure such as intro, verse, chorus, and bridge
- Designed for professional music production and advanced creative tools
Lyria 3 Clip (lyria-3-clip-preview)
- Produces 30-second clips optimized for speed
- Ideal for prototyping, background loops, and social media content
Both versions support realistic AI-generated vocals with improved clarity and expressive detail.
Granular controls and multimodal inputs
Lyria 3 introduces several controls to help creators guide music generation more precisely:
- Tempo conditioning to set the rhythm and pace of the track
- Time-aligned lyrics to control when lyrics appear in the composition
- Image-to-music input, allowing users to provide an image that influences the mood and style of the music
Integration across Google platforms
Google has integrated Lyria 3 Pro into multiple products and developer tools:
- Vertex AI: scalable music generation for business applications like game soundtracks and creative tools
- Google AI Studio and Gemini API: enabling developers to build AI music apps
- Google Vids: generating custom music for marketing or personal videos
- Gemini app: allowing paid users to create longer music tracks for vlogs, podcasts, and tutorials
- ProducerAI: collaborative AI songwriting and music production
Example use cases
Demonstrations in AI Studio show several potential applications:
- Automatically generating background music for videos
- Creating personalized songs for alarm clock apps based on weather or calendar data
- Composer mode to build songs section by section
- Text mode to generate music using natural language prompts
Responsible AI safeguards
Google says the system includes measures to ensure responsible use of AI music:
- SynthID watermarking embedded in generated tracks to identify AI-created audio
- Training based on authorized materials from YouTube, partner agreements, and applicable laws
- Filters preventing the model from imitating specific artists or violating intellectual property
The technology has also been tested by professional musicians including Grammy-winning producer Yung Spielburg and DJ François K, who used it in music production and film scoring projects.
Lyria 3 and Lyria 3 Pro Availability
- Lyria 3 Clip and Lyria 3 Pro are available globally in public preview for developers
- Accessible through Gemini API and Google AI Studio
- Integrated with tools such as Google Vids, Gemini app, and ProducerAI
- Documentation, demo apps, and code samples are available to help developers integrate the models.