Google just dropped Lyria 3, its most advanced AI music generation model yet. The system can produce full 30-second instrumental tracks from text prompts, and it's surprisingly good.
Currently in beta for US users through YouTube's Dream Track feature, Lyria 3 represents a significant leap from its predecessor. The model handles complex arrangements across genres—from orchestral compositions to electronic beats—with notably improved audio fidelity and musical coherence.
The timing matters. AI-generated music has been a legal and creative minefield, with major labels suing generators left and right. Google's approach threads the needle by partnering with labels rather than fighting them, offering revenue sharing and attribution systems built into the platform.
For creators, this means high-quality background music for YouTube videos without licensing headaches. For the music industry, it's another signal that AI composition tools are moving from novelty to utility.
The 30-second cap is clearly intentional—long enough to be useful, short enough to avoid replacing full songs. Expect that limit to expand as licensing deals mature.
Organizing the world’s information and making it accessible through search
private
1600 Amphitheatre Parkway Mountain View, CA 94043, USA
1998