Google's Lyria 3 Pro, the latest iteration of its AI music model, now generates complete three-minute songs—a significant leap from the 30-second clips launched last month. This update allows users to craft detailed compositions by specifying elements like intros, verses, choruses, and bridges, leveraging improved musical understanding for complex transitions and stylistic experimentation.**

The tool is available to paid Gemini users and enterprise clients via Vertex AI, with developer access through the Gemini API and Google AI Studio. It’s also integrated into Google Vids, an AI video platform, enabling seamless audio-visual creation. Notably, Lyria 3 Pro prioritizes ethical AI use, training exclusively on licensed materials and embedding SynthID watermarks to identify AI-generated content—addressing industry concerns about authenticity.

The launch coincides with Spotify’s ongoing battle against AI-generated spam: approximately 50,000 AI tracks are uploaded daily, prompting the platform to delete 75 million tracks last year. This underscores the tension between innovation and quality control in AI music.**

While Lyria 3 Pro offers creative flexibility, its success hinges on balancing artistic potential with responsible deployment. For now, it remains a niche tool for experimentation, leaving broader adoption to platforms grappling with scale and authenticity.