Google Launches Lyria 3 Pro Music Generation Model

Google has officially launched Lyria 3 Pro, its advanced music generation model, offering users the ability to create full-length tracks up to three minutes in duration with enhanced structural control. The model represents a significant upgrade from the previous Lyria 3 release, which limited tracks to 30 seconds, and introduces capabilities for specifying musical elements like intros, verses, choruses, and bridges through natural language prompts.

The Google's official Lyria 3 Pro announcement details that the model integrates seamlessly into existing Google products, including the Gemini app (available to paid subscribers), Google Vids video editor, and ProducerAI, a music creation tool acquired by Google last month. Enterprise users can leverage Lyria 3 Pro via Vertex AI (in public preview), the Gemini API, and AI Studio, enabling scalable music production for applications ranging from gaming soundtracks to marketing videos.

Unlike its predecessor, Lyria 3 Pro demonstrates improved musical composition understanding, allowing users to generate tracks with complex arrangements while maintaining high fidelity. Google emphasized that the model was trained using permissible data from YouTube and Google's own content, with all generated tracks marked by SynthID watermarking to indicate AI creation. The company clarified that while users can specify artist styles in prompts, the model takes "broad inspiration" rather than mimicking specific artists.

Industry context highlights Google's strategic positioning against competitors like Spotify and Deezer, which recently introduced tools to help artists identify AI-generated music and prevent misattribution. Google's approach aligns with broader industry efforts to address copyright concerns, though the company has not yet faced direct legal challenges related to its music generation training data.

Artist partnerships underscore the model's creative potential. Grammy-winning producer Yung Spielburg used Lyria 3 in composing the score for Google DeepMind's short film "Dear Upstairs Neighbors," while DJ Françoise K collaborated on a song using Lyria 3 Pro's iterative capabilities. Spielburg noted the model's ability to "move beyond the limits of traditional instrumentation," creating "technically impossible" sonic timbres.

Enterprise adoption is already underway, with companies like Artlist and Freepik integrating Lyria 3 models into their workflows. Artlist's Chief Product Officer noted the model's "unprecedented level of creative control and high-fidelity output," while Freepik's engineering lead highlighted reduced iteration time and "more predictable" music generation for real-world creative pipelines.

Technical specifications reveal Lyria 3 Pro generates 48kHz stereo audio from text prompts or image inputs, with support for vocal generation, timed lyrics, and multi-modal composition. The model's 131,072 token input limit accommodates detailed structural instructions, such as specifying "a 3-minute track with a 15-second intro, two verses, a chorus, and a bridge." Google Cloud's documentation confirms the model's integration into Vertex AI for businesses requiring on-demand audio at scale, enabling applications from gaming soundtracks to creative tool integrations.

Artūras Malašauskas is an AI Systems Integrator with 20+ years of production-grade web engineering experience. He has designed, shipped, and scaled enterprise Python/PHP systems for logistics, SaaS, and public-sector clients. For the past year, he has focused exclusively on AI integrations: deploying open-source LLMs, building generative media pipelines (image, audio, video), and engineering multi-agent workflows for real production environments. His standard: reproducibility, security, cost-efficient inference—no vaporware. He documents and evaluates emerging AI tooling, separating verified capabilities from marketing noise. Technical editor at: muza-ai.eu, ai-verslas.lt, ai-naujinos.lt Connect on LinkedIn