Google has unveiled Lyria 3 and Lyria 3 Pro, its latest music generation models designed to empower developers with advanced creative capabilities. The models are now rolling out in public preview through the Gemini API and a new dedicated audio experience within Google AI Studio.
The new generation of Lyria represents a significant step forward in AI-powered music creation, combining sophisticated musical awareness with structural coherence. This enables developers to build applications that produce high-fidelity compositions complete with realistic vocals, verses, and choruses while maintaining musical consistency throughout entire tracks.
Google offers two distinct model variants tailored to different production needs. Lyria 3 Pro serves as the premier option for comprehensive song generation, creating full-length tracks up to approximately three minutes with professional-grade structural awareness suitable for studio-quality output. For developers prioritizing speed and scalability, Lyria 3 Clip generates high-quality 30-second clips optimized for rapid prototyping, background loops, and social media content creation.
Both variants feature realistic vocal generation with expressive nuance and improved clarity for natural-sounding audio. The models support a wide range of global languages and musical genres, from pop and funk to Motown, giving creators substantial flexibility in artistic direction.
A standout feature of Lyria 3 is its precision control system. Developers can direct the model through natural language prompts with granular control over tempo conditioning, allowing them to set specific rhythms with high accuracy. The platform supports time-aligned lyrics, enabling developers to outline song progression and control precisely when lyrics begin and end. Additionally, Lyria 3 introduces multimodal image-to-music input, allowing developers to provide images that influence the mood, style, and atmospheric qualities of generated audio.
Google has created several demonstration applications within AI Studio showcasing practical use cases. One example enables users to upload videos that are analyzed to generate descriptive prompts, which Lyria then transforms into synchronized background music. Another demo application generates personalized alarm clock songs containing relevant information such as weather, location, time, date, and calendar events.
The dedicated music generation workspace in Google AI Studio provides developers with a first-class environment to experiment with Lyria 3's advanced features. Using a paid API key, developers can access the playground and explore powerful creation modes to bring their musical visions to life.