Google DeepMind Launches Lyria 3 Pro for Structured AI Music

TL;DR: Google DeepMind has launched Lyria 3 Pro, an AI music generation model capable of producing full-length tracks up to three minutes long. Unlike previous iterations limited to 30-second clips, the new model features structural awareness, allowing users to explicitly prompt for specific song components like verses, choruses, and bridges. The model is now integrated into the Gemini app, Google Vids,

Key facts

• Lyria 3 Pro can generate music tracks up to three minutes long, a significant increase from previous versions limited to 30-second clips [1, 2, 3, 4, 5].
• The model features ‘structural awareness,’ allowing users to prompt for specific elements like intros, verses, choruses, and bridges [1, 2, 3, 4, 5].
• Lyria 3 Pro generates full songs including vocals, lyrics, and instrumentation [1, 5].
• The model is developed by Google DeepMind [1, 5].
• Lyria 3 Pro is integrated into the Gemini app for paid subscribers [2, 4].
• The model is available in Google Vids, rolling out to Google Workspace and Google AI Pro & Ultra subscribers [3, 4].
• Enterprise access is provided through Vertex AI, currently in public preview [2, 3, 4].

Google DeepMind Unveils Lyria 3 Pro: A Shift to Structured, Full-Length AI Music

Google DeepMind has officially launched Lyria 3 Pro, a significant advancement in its AI music generation capabilities. This new model marks a pivotal shift from the company’s previous iterations, which were restricted to generating short, 30-second audio clips [2, 3]. Lyria 3 Pro is designed to produce full-length music tracks up to three minutes long, complete with vocals, lyrics, and complex instrumentation [1, 5].

The introduction of Lyria 3 Pro positions Google to directly challenge established royalty-free music services like Epidemic Sound, particularly for content creators who need personalized tracks for vlogs, podcasts, and tutorials [3]. By offering longer, more structurally coherent songs, Google is expanding the utility of AI-generated music from background ambience to primary audio content.

Structural Awareness: Defining the Song’s Architecture

A defining feature of Lyria 3 Pro is its “structural awareness.” Unlike earlier models that relied on the AI to infer the arrangement of a song, this update allows users to explicitly prompt for specific song components [1, 2, 3, 4, 5]. Users can now direct the model to generate distinct sections such as intros, verses, choruses, bridges, and outros [1, 2, 3, 4, 5].

This level of control enables more complex musical compositions that mimic the structure of traditional songs. The model generates complete songs that include vocals and lyrics, moving beyond instrumental loops to full musical pieces [1, 5]. This capability is critical for creators who need music that aligns with specific narrative beats or emotional arcs in their content.

Integration Across Google’s Ecosystem

Google is rolling out Lyria 3 Pro across multiple platforms to serve diverse user bases, from casual creators to enterprise developers.

Consumer and Creator Access

For general users, Lyria 3 Pro is available in the Gemini app for paid subscribers [2, 4]. This integration allows users to generate music directly within Google’s AI assistant interface. Additionally, the model is being integrated into Google Vids, a video editing tool, and is rolling out to Google Workspace and Google AI Pro & Ultra subscribers [3, 4].

Google is also integrating Lyria 3 Pro into ProducerAI, a platform recently acquired by Google, further expanding its reach in the music creation space [2]. These integrations suggest a strategy to embed AI music generation seamlessly into the workflows of video creators and podcasters.

Enterprise and Developer Access

For enterprise clients and developers, Lyria 3 Pro is accessible via Vertex AI, which is currently in public preview [2, 3, 4]. This allows businesses to integrate the model into their own applications and services. Additionally, developers can access the model through Google AI Studio and the Gemini API [2, 3, 4].

This broad availability across consumer and enterprise platforms highlights Google’s commitment to making AI music generation a core feature of its cloud and productivity ecosystem.

Safety, Copyright, and Ethical Considerations

Google has addressed potential concerns regarding copyright and artist mimicry. The company states that Lyria 3 Pro does not mimic specific artists [2]. When users include named creators in their prompts, the model treats them as broad inspiration rather than direct imitations [2]. This approach aims to balance creative freedom with respect for existing artists’ work.

To prevent infringement, Google employs checks against existing content [2]. Additionally, all outputs from Lyria 3 Pro include a silent SynthID watermark to identify them as AI-generated [2]. This watermarking is part of Google’s broader effort to promote transparency in AI-generated media.

Market Implications

The launch of Lyria 3 Pro comes at a time when AI music generation is becoming increasingly competitive. Services like Suno and Udio have already gained traction for their ability to generate high-quality, full-length songs [3]. By offering a model that can produce three-minute tracks with structural control, Google is positioning Lyria 3 Pro as a viable alternative for professional and semi-professional creators.

The integration of Lyria 3 Pro into Google’s existing products, such as Gemini and Vids, provides a significant advantage in terms of accessibility and ease of use. For creators who are already within the Google ecosystem, the ability to generate music directly within their workflow could make Lyria 3 Pro a preferred choice over third-party services.

Conclusion

Lyria 3 Pro represents a major step forward in AI music generation, offering longer, more structured, and more controllable musical outputs. Its integration across Google’s consumer and enterprise platforms suggests a strategic move to dominate the AI music generation market. As the technology continues to evolve, it will be interesting to see how Google balances innovation with ethical considerations and how it competes with established players in the industry.

The model was announced as part of Google’s AI push at I/O 2025, signaling its importance in the company’s broader AI strategy [5]. With its focus on structural awareness and broad accessibility, Lyria 3 Pro is poised to impact the way music is created and consumed in the digital age.

Sources

3. Watch out Epidemic Sound: Google launches Lyria 3 Pro AI model that can generate 3-minute tracks - Music Business Worldwide (www.musicbusinessworldwide.com) — 2026-03-26
4. Google Lyria 3 Pro makes longer AI songs (www.theverge.com) — 2026-03-25
5. What Is Google Lyria 3 Pro? How to Generate Full-Length AI Music with Structural Control (www.mindstudio.ai) — 2026-03-30