Google Lyria 3 Pro: AI Music Generation in 2026

TL;DR

• Google launched Lyria 3 Pro on March 25, 2026, upgrading from 30-second clips to full 3-minute tracks with structural awareness of intros, verses, choruses, and bridges.¹
• The model is available across six platforms: the Gemini app (paid subscribers), Google Vids, ProducerAI, Vertex AI (public preview), the Gemini API, and AI Studio.¹²
• Every generated track carries a SynthID watermark that survives common modifications like MP3 compression and noise addition, ensuring AI-generated music remains identifiable across distribution channels.²
• Google claims Lyria 3 Pro was trained on partner-licensed data and permissible content from YouTube, positioning it as a legally safer alternative to Suno and Udio, both of which faced copyright lawsuits from major record labels in 2024–2025.³⁴
• The generative AI in music market was valued at approximately $570 million in 2024 and is projected to reach $2.8 billion by 2030, growing at a CAGR of 30.5%.⁵

What You'll Learn

In this article, you will learn how Google Lyria 3 Pro works, what makes it different from the original Lyria 3, how it stacks up against Suno v5.5 and Udio in quality and features, what the enterprise API access means for developers, and how SynthID watermarking addresses the legal minefield around AI-generated music.

What Is Google Lyria 3 Pro?

Google Lyria 3 Pro is an AI music generation model developed by Google DeepMind. It is the successor to Lyria 3, which launched just one month earlier in February 2026 with a 30-second generation limit. The "Pro" upgrade extends that cap to three full minutes and introduces structure-aware composition — the model understands musical elements like intros, verses, choruses, and bridges, and users can specify these in their prompts.¹

The model sits within Google's broader generative AI ecosystem. Rather than offering it as a standalone product, Google has distributed Lyria 3 Pro across six different surfaces: the Gemini app for consumers, Google Vids for video editors, ProducerAI for music producers, and Vertex AI, the Gemini API, and AI Studio for developers and enterprises.¹²

This multi-platform distribution strategy is significant. While Suno offers enterprise API access through its Enterprise tier, Google's deep integration across consumer apps, developer tools, and enterprise infrastructure gives it a structural advantage in the B2B segment of AI music generation. The Vertex AI integration is powered by Google's custom TPU infrastructure — for context on how Google and other tech giants are building custom silicon to run models like Lyria 3 Pro, see our breakdown of the custom AI chip race in 2026.

How Lyria 3 Pro Compares to Its Predecessor

The jump from Lyria 3 to Lyria 3 Pro is substantial for a one-month release gap. Here is what changed:

The pricing for Lyria 3 Pro on Vertex AI has not been officially disclosed as of March 27, 2026. For reference, the earlier Lyria 2 model was priced at $0.06 per 30 seconds of generated audio on Vertex AI.⁶

The AI Music Generation Landscape in 2026

Google is not entering an empty market. The AI music generation space has consolidated around three major players, each with distinct strengths and legal positions.

Suno v5.5

Suno released version 5.5 on March 26, 2026 — one day after Google's Lyria 3 Pro announcement — adding three major features: Voices (use your own verified singing voice for Pro and Premier subscribers), Custom Models (fine-tune v5.5 on your own music, up to three per account), and My Taste (a personalization engine).⁷

Suno v5, released in September 2025, already offered tracks up to 8 minutes long with what the company calls "Intelligent Composition Architecture." In Suno's own benchmark testing, v5 achieved an ELO score of 1,293, placing it ahead of competitors in audio fidelity, musical structure, and vocal realism.⁸ Suno also provides a virtual DAW (digital audio workstation) interface and stem export for professional workflows.

Udio

Udio's position in 2026 is in transition. In October 2025, Udio settled its copyright infringement lawsuit with Universal Music Group and announced a strategic partnership to build a new licensed AI music creation platform.⁹ The platform has been transitioning its training data to UMG's licensed catalog. During this transition, all audio, video, and stem downloads are temporarily disabled across plans.⁹ New features include multi-language vocal generation and stem separation, but the download restrictions have frustrated users who relied on the platform for production work.

How They Compare

The comparison reveals that no single platform dominates across all dimensions. Suno leads in audio quality and creator tools. Google leads in enterprise distribution and legal defensibility. Udio is rebuilding around a licensed model but currently has the weakest user experience due to download restrictions.

The Copyright Question

The legal landscape around AI music generation is the undercurrent shaping every product decision in this space.

In June 2024, the Recording Industry Association of America (RIAA) and major record labels including Universal Music Group, Sony Music, and Warner Music filed copyright infringement lawsuits against both Suno and Udio, alleging that the platforms trained on copyrighted recordings without permission.⁴ Udio has since settled with UMG (October 2025) and Warner Music (November 2025), but Sony Music's case against Udio remains active. The Suno lawsuit is still working through the courts as of March 2026.

Google's approach with Lyria 3 Pro is explicitly designed to sidestep this risk. The company states that it trained the model using data from its partners and "permissible data from YouTube and Google" properties.³ When users reference a specific artist in their prompts, the model takes "broad inspiration" rather than attempting to replicate that artist's sound directly.

Additionally, every track generated by Lyria 3 or Lyria 3 Pro includes a SynthID watermark — a digital fingerprint developed by Google DeepMind that is embedded in the audio signal itself and survives common transformations like MP3 compression, noise addition, and speed changes.² This makes it possible to trace AI-generated music even after it has been distributed and modified.

Udio has taken a different but parallel path, transitioning to Universal Music Group's licensed catalog for training data. Under this new model, each AI-assisted creation contributes to a royalty pool distributed to the licensed UMG artists whose work informed the training data.⁹

Suno has not publicly announced a similar licensing arrangement and continues to face the pending RIAA lawsuit. Suno is also facing a separate copyright case in Germany brought by GEMA (the German music rights organization), with a ruling scheduled for June 2026.¹⁰

Enterprise and Developer Access

For developers and businesses, the most significant aspect of Google's announcement is the Vertex AI integration. Lyria 3 Pro is available in public preview on Vertex AI, with access also available through the Gemini API and AI Studio.¹

This opens several use cases that were previously difficult to address at scale: gaming studios generating dynamic soundtracks that adapt to gameplay, video platforms offering automated background music for creator content, advertising agencies producing custom jingles without licensing fees, and podcast platforms generating intro and outro music. These enterprise use cases overlap with the broader trend of AI-powered content creation tools that are reshaping how businesses produce media at scale.

Google is explicitly targeting gaming studios and video platforms with the Vertex AI integration, positioning Lyria 3 Pro as infrastructure rather than a consumer toy.⁶

While Suno offers API access through its Enterprise tier and Udio provides limited API access on Pro and Enterprise plans, neither matches the breadth of Google's integration across Vertex AI, the Gemini API, AI Studio, and consumer products. Google's advantage lies in the seamless connection between its music generation model and its existing cloud and productivity ecosystem, making it the most straightforward option for developers already building on Google Cloud.

What This Means for Creators and Developers

The practical takeaway depends on your use case.

If you are a music creator looking for the best AI-assisted composition tool, Suno v5.5 remains the strongest option. Its vocal quality, custom model fine-tuning, personal voice synthesis, and stem export features are designed for production workflows. Lyria 3 Pro's lack of personal voice synthesis and stem export makes it less suitable for serious music production.

If you are a developer building an application that needs programmatic music generation, Lyria 3 Pro on Vertex AI is currently the most accessible enterprise option. The API access, SynthID watermarking, and Google's licensed training data provide a more defensible foundation for commercial applications.

If you are a content creator who needs background music for videos, podcasts, or social media, Lyria 3 Pro through the Gemini app offers the most convenient experience — assuming you already have a paid Gemini subscription. The integration with Google Vids makes it particularly useful for video editors in the Google ecosystem.

If legal risk is your primary concern, Google's licensed training data and SynthID watermarking offer the strongest position as of March 2026, though Udio's UMG partnership is a close second once its download restrictions are lifted.

Footnotes