Google brings stronger AI music creation to Flow

Google is rolling out a stronger Lyria music model inside Flow Music, aiming to make AI-generated songs sound more natural and easier to shape.

The practical change is simple: people using Flow Music should be able to create richer tracks with more complex melodies, better prompt-following lyrics, and vocals that carry more expression. Google says pronunciation is also improved, which matters for generated songs because small vocal flaws quickly make synthetic music feel uncanny.

The update also adds more direct creative control. Users can steer tempo and duration more easily instead of treating the model like a one-shot song machine. That pushes Flow Music closer to an interactive creative tool, where the useful part is iteration: trying a mood, adjusting the shape, and getting something shareable without a full production workflow.

Google frames the launch as part of its broader Lyria family from DeepMind. For general users, the important bit is not the model label. It is that high-quality music generation is moving into everyday Google creative tools, alongside images and video, with controls that make the output less random and more personal.