We’re launching Lyria 3.5 in Google Flow Music, with advances across musicality, lyrics, vocals, and creative…

Google is launching Lyria 3.5 in Google Flow Music. The company describes Lyria 3.5 as its newest music generation model, with advances across musicality, lyrics, vocals, and creative control, aimed at letting people craft richer tracks.

Lyria 3.5 sits inside Google Flow Music rather than as a standalone research demo. The model targets four product-facing dimensions at once: musicality, lyrics, vocals, and creative control. That framing points to an end-to-end generation stack where composition quality, text-to-lyric coherence, vocal realism, and user-directed steering are improved together instead of one isolated capability.

For engineers and builders, the signal is that music generation is being productized as a controllable creative tool, not only a one-shot prompt-to-audio experiment. If you ship audio features, agentic creative workflows, or content tooling, the relevant bar is multi-axis quality—structure and feel, lyric fidelity, vocal character, and how much the user can steer—because that is what Google is shipping into Flow Music.

In market terms, Google is putting a named music model into a consumer creative surface and staking the release on breadth of improvement rather than a single headline metric. Competitors in generative audio will be measured against the same bundle: musicality, lyrics, vocals, and control. A model that only wins on one of those axes will look incomplete next to a Flow Music integration that claims gains across all four.

Watch how Flow Music exposes creative control in the product UI—what parameters users can lock, edit, or iterate on after a first generation—and whether Lyria 3.5’s musicality, lyric, and vocal gains show up as consistently usable tracks under real prompts. The next practical test is whether builders adopt Flow Music as a production step in music workflows or treat it as a drafting tool only.