Google's Lyria 3 Pro can now generate AI music (slop) up to 3 minutes in length
TL;DR
Google released Lyria 3 Pro, upgrading its AI music model to generate full songs up to 3 minutes long – up from just 30 seconds at launch last month.
Key Points
- Users can now prompt specific song elements like intros, verses, choruses, and bridges for more structured compositions.
- Google claims improved understanding of musical composition and better handling of complex style transitions.
- Available now for paid Gemini users, enterprise customers on Vertex AI, and developers via the Gemini API and Google AI Studio.
- The model is also being integrated into Google Vids, the company's AI video generation platform.
Nauti's Take
The 'slop' label in the Engadget headline is more honest than anything in Google's press materials. Lyria 3 Pro generates music faster and longer, but the real question remains unanswered: does it sound good enough to displace actual music production, or is it just better elevator music?
The structural prompting is a genuine step forward, but 'better understanding of musical composition' is classic benchmarkless marketing. What's more interesting than the model itself is the distribution play – anyone already paying for Gemini or using Vertex AI gets this automatically, which is where the real adoption leverage lies.
Context
The jump from 30 seconds to 3 minutes isn't just a technical milestone – it's what makes the tool actually usable in real production contexts. Anyone needing background music for videos, podcasts, or ads can now generate complete tracks without stitching clips together. The ability to prompt specific structural elements like verse and chorus gives non-musicians a meaningful layer of compositional control for the first time.
Integrating Lyria into Vertex AI and Google Vids simultaneously signals that Google is positioning this as infrastructure, not a demo.
Video
Sources