Gemini gets a major upgrade as Google adds 3-minute AI music generation: here’s how it works

Google is baking yet another powerful capability into Gemini with the integration of longer music generation in the AI chatbot, thanks to the Lyria 3 Pro integration. While Google had integrated its music generation model earlier in the year into Gemini, the integration of Lyria 3 Pro now allows users to create up to 3-minute-long tracks, a big upgrade from the original model, which could only generate 30-second audio snippets.

What’s new with Lyria 3 Pro?

Apart from generating 3-minute-long tracks, Google has added many other new abilities to its music generation model. With Lyria 3 Pro, Gemini can now create higher-fidelity tracks with verses, choruses, and bridges. The new upgrade also allows users to generate long instrumental backing tracks for studying, custom soundtracks for YouTube vlogs and social posts, or even original jingles for small businesses and personal brands.

“This advanced version allows the creation of tracks up to 3 minutes long, with customization and creative control. Lyria 3 Pro better understands musical composition, so you can now prompt for specific elements like intros, verses, choruses, and bridges,” the company said in a blog post.

Google says since Lyria 3 Pro better understands musical composition, it’s great for experimenting with different styles or generating songs with complex transitions.

Apart from the Gemini app, Google is also adding music generation capabilities to various other products, including Google Vids, ProducerAI, Vertex AI, ProducerAI, and AI Studio.

Google says it has developed Lyria 3 Pro in collaboration with the music industry. The company also says that Grammy-winning producer Yung Spielburg had also recently used Lyria in his composition process for a short film score, while renowned DJ François K used it to iterate on a soon-to-be-released song.

The company says that if you ask Gemini to create a song mimicking a specific, real-world artist, it will only take broad creative inspiration to match the mood rather than directly copying proprietary voice or style.

“We employ filters to check outputs against existing content, and users must adhere to the Terms of Service and Gen AI prohibited use policies, which prohibit violating others’ intellectual property and privacy rights,” the company said in a blog post.

Meanwhile, Google also says that all tracks generated with Lyria 3 and Lyria 3 Pro are embedded with SynthID, the company’s watermark. Meanwhile, users can also upload a suspicious audio file to Gemini to check for the SynthID watermark and verify its origin.