Introducing Lyria 3.5 Music Generation in Gemini

Summary
This official promo video from Google introduces Lyria 3.5 music generation integrated into the Google Gemini app. It demonstrates the new music creation interface, including vocal controls, genre options, longer track durations, and starter templates.

What is shown

• [00:01] Opening title card: "Turn it up with Lyria 3.5".
• [00:05] Gemini music interface showing options for Length, Vocals ("Vocals on" / "Instrumental"), and Genre.
• [00:10] Playback demonstration of an AI-generated K-pop track titled "New Orbit" with Korean/English vocals and animated album art.
• [00:15] Genre picker menu displaying options such as Pop, Hip-hop & rap, Rock, K-pop, Afrobeats, Electronic, R&B, Country, Latin, and Reggae; selecting "Latin".
• [00:20] Playback of a Latin vocal track titled "Quince Vueltas".
• [00:25] Track length configuration setting showing "Short" and "Standard" options, selecting "Standard" which extends tracks up to 3:00 minutes [00:30].
• [00:32] Playback sample of a Soul track titled "Our Atlas".
• [00:37] Template browser showing starter presets (e.g., "Intro music", "Say it with a track", "Rap battle", "Speech to song", "Business promo", "Streaming", "Brand jingle", "Soundscape", "Birthday roast"), previewing the "Business promo" audio track.
• [00:44] Ending title card: "That's Google Gemini", directing users to get the Gemini AI app.

Claims & numbers

• The video displays track lengths running up to 3:00 minutes ("00:15 / 03:00") [00:31].
• Supports customizable vocal mode toggles ("Vocals on" or "Instrumental") [00:08].
• Displays 10+ genres and a wide selection of preset creation templates [00:17, 00:38].
• Disclaimer notes: "Limits apply. Compatibility varies. 18+" [00:46].

Notable quotes

• [00:01] "Turn it up with Lyria 3.5" (on-screen text)
• [00:03] "Cue the voice or leave it out" (on-screen text)
• [00:23] "Let your tracks run longer" (on-screen text)

Assessment
This is an official commercial feature showcase by Google Gemini using animated UI mockups and generated audio samples. While the workflow demonstrates actual feature settings (length, vocals, genres, and templates), the full prompt generation cycle is presented in edited, stylized snippets rather than an uncut, end-to-end interface recording.

Described by gemini-3.8-flash on 2026-10-06 from the video's audio and frames.