Google just dropped Lyria 3 Pro, a significant upgrade to its music generation model that pushes AI-created tracks from 30 seconds to a full three minutes. The release, reported by TechCrunch AI, comes just one month after the original Lyria 3 launched inside the Gemini app.
The jump from half-a-minute clips to three-minute tracks is the headline feature, but it’s not the only one worth noting.
🎵 What Lyria 3 Pro Can Do
- Longer tracks: Up to 3 minutes, a 6x increase over Lyria 3’s 30-second limit
- Structural awareness: The model understands intros, verses, choruses, and bridges, so users can specify song structure directly in their prompts
- Better creative control: More customization options for shaping the output beyond a simple text description
- Artist-inspired generation: Users can reference an artist’s style. Google says the model takes “broad inspiration” rather than mimicking anyone directly
🔑 Who Gets Access
This isn’t a freebie. Lyria 3 Pro is rolling out inside the Gemini app, but only paid subscribers can use it. That’s a clear signal Google sees premium AI music generation as a feature worth gating.
Beyond consumer access, Google is pushing Lyria 3 Pro into its business and developer tools:
- Google Vids (video editing app)
- ProducerAI (the GenAI music production tool Google acquired last month)
- Vertex AI (public preview for enterprise customers)
- Gemini API and AI Studio (for developers building on top of it)
The enterprise play is interesting. By making Lyria 3 Pro available through Vertex AI and the Gemini API, Google is positioning itself to capture the commercial music generation market, think background tracks for ads, videos, and apps, not just consumer experimentation.
🛡️ The Trust and Safety Angle
Google is being careful about the provenance question. According to TechCrunch AI, the company emphasized it trained the model on partner data and “permissible data” from YouTube and Google. Every track generated by Lyria 3 or Lyria 3 Pro gets watermarked with SynthID, Google’s AI content labeling system.
This matters because the music industry is watching AI generation closely. Spotify just released tools letting artists review songs published under their name to fight AI misattribution. Deezer launched detection tools that any streaming service can use to flag AI-generated music. Google watermarking everything by default is a smart preemptive move.
📌 Why This Matters
The music generation space is heating up fast. Suno, Udio, and others have been pushing boundaries, but Google bringing this capability directly into Gemini, its API ecosystem, and enterprise tools changes the distribution game entirely. Millions of Gemini subscribers now have a built-in music creation tool. Developers can integrate it into their apps through familiar Google APIs.
The structural understanding feature is what stands out most. Moving from “generate a song about X” to “generate a song with a slow intro, two verses, a chorus, and a bridge” is a meaningful step toward making these tools actually useful for people who know what they want.
For more details, check the original reporting from TechCrunch AI.