Google Launches Lyria 3.5 Music Generation Model in Flow Music

Google’s newest AI maestro, Lyria 3.5, has just hit the digital studio floor. On July 29 2026, the company rolled out the model to users of Google Flow Music, its web‑ and mobile‑based platform that lets creators write, remix, and share full songs.

Lyria 3.5 follows a lineage of DeepMind‑powered music models that have steadily expanded the toolkit available to musicians. Earlier releases—Lyria 2, Lyria 3, and Lyria 3 Pro—introduced the ability to generate instrumental tracks, lyrics, and vocal lines from simple text prompts. According to Google, Lyria 3.5 builds on those foundations by delivering richer, more complex melodic structures, higher‑quality lyric writing, and vocal performances that sound noticeably more human.

The new model offers tighter control over tempo and track duration, allowing users to shape a song’s architecture with greater precision. Users can now dictate the length of verses, the pace of choruses, and the timing of bridges, giving the AI a clearer blueprint for the final piece.

Technical details released by Google underscore the model’s leap in audio fidelity. Vocal output shows improved pronunciation and emotional nuance, while the lyric engine follows prompts with a higher degree of accuracy, whether the instructions are straightforward or elaborate. The updated architecture also enhances the model’s awareness of song structure, helping it maintain coherence across verses, choruses, and bridges.

Lyria 3.5’s debut comes after a flurry of AI announcements from Google over the past five months. Earlier this week, the company unveiled Gemini 3.6 Flash, a high‑throughput language model, and its variants Gemini 3.5 Flash‑Lite and Gemini 3.5 Flash Cyber. Lyria 3.5 is DeepMind’s third major music‑generation model in that period, underscoring the firm’s focus on expanding AI‑driven creative tools.

Google Flow Music is positioned as a professional‑grade studio. It lets users generate original tracks from text or image prompts, remix existing songs, and download royalty‑free audio. The platform offers a free tier and a subscription plan that starts at $6 per month. With Lyria 3.5 now embedded in Flow Music, creators can access the model’s enhanced melodic and vocal capabilities directly within the same interface.

Industry observers note that AI‑generated music is gaining traction across streaming services, advertising, and content creation. By adding Lyria 3.5 to Flow Music, Google expands its footprint in this growing segment. The company’s approach differs from some competitors in that it embeds the model within a broader ecosystem that includes Google’s cloud services, search, and productivity tools.

At the time of the announcement, Lyria 3.5 is live and available to any Flow Music user. Google has not disclosed plans for additional updates or new features beyond the current release. The model’s availability marks a continued effort by DeepMind to provide tools that support musicians, producers, and hobbyists in creating high‑quality audio content.

In short, Google’s Lyria 3.5 delivers significant improvements in melody, lyric generation, vocal realism, and creative control. Its integration into Flow Music gives users a ready‑to‑use platform for experimenting with AI‑generated songs, reflecting the company’s broader strategy to advance AI tools for creative industries.