Google Flow Music Launches Lyria 3.5 With Full-Length Songs and Music Videos

Google DeepMind launched Lyria 3.5, its latest AI music generation model, rolling out today inside Google Flow Music, the company's dedicated music-creation platform.

The upgrade focuses on three things: more expressive, dynamic vocals; richer musical arrangements that flow more naturally; and tighter creative control, including BPM adjustment and exportable stems for full-length songs. It's a mid-cycle refinement rather than a full rebuild, following Lyria 3's February debut and Lyria 3 Pro's March expansion to three-minute tracks.

What Actually Changed

Google's own framing breaks the upgrade into four areas. Musicality is described as producing richer, more complex melodic structures. Lyrics are said to follow prompts more accurately while maintaining clearer song structure. Vocals get the most emphasis, with Google promising more realistic, emotionally nuanced singing and improved pronunciation. And creative control now extends to specifying tempo and duration directly, rather than leaving structure entirely up to the model.

Google hasn't said whether Lyria 3.5 replaces Lyria 3 Pro across the Gemini app and developer platforms, or whether it's specific to Flow Music for now.

Flow Music Is More Than a Song Generator

Flow Music's pitch extends past text-to-song generation. The platform includes a "Producer" chat interface for iterating on tracks conversationally, direct integration with Google's Veo video model for generating accompanying music videos, a "vibe-coding" space for building custom instruments and audio tools, and stem-splitting and remix tools for reworking generated tracks after the fact.

The Veo integration is the detail most relevant to BRC's beat specifically: Flow Music lets users direct a music video for a generated song using the same video model already used for AI filmmaking elsewhere, controlling characters, aesthetics, and visual details without a production crew. That's a genuinely different pitch than a standalone music generator — it's positioning Flow Music as one node in Google's broader Flow creative ecosystem, alongside Imagen for stills and Veo for video.

Access and Pricing

Flow Music is free to start, with daily generation credits and no credit card required. Paid access maps to existing Gemini subscription tiers: Google AI Plus members get Flow Music's Starter plan, AI Pro members get the Plus plan, and AI Ultra members get the Member plan. All outputs carry Google's SynthID watermark for AI-content identification.

Competitive Context

Flow Music enters a crowded AI music space that includes Suno and Udio as the most established consumer-facing names, alongside newer entrants building similar text-to-song tools.

Google's differentiation is less about raw output quality and more about ecosystem integration — tying music generation directly to Veo for video and folding pricing into existing Gemini subscription tiers rather than a standalone product.

The Signal in the Noise

For filmmakers and video creators, the Veo-integrated music video feature is the part worth paying closer attention to, more than the vocal quality upgrade itself. A tool that generates a song and then directs a matching AI music video from the same platform collapses a workflow that currently requires stitching together separate music and video generation tools.

Whether the video output quality holds up under real use is worth testing directly, but the integration itself is a meaningful step toward single-platform AI content production rather than a chain of specialized tools.

Resources & Reads

• We're Launching Lyria 3.5 — Official Google Blog
• Google's Lyria 3.5 Makes AI Songs More Expressive — Coverage on Neowin
• Google Flow Music