Google has upgraded its Gemini with the new Lyria 3 Pro music-generation model, offering more advanced song creation features.
Lyria 3 Pro improves understanding of music structure and supports tracks up to three minutes long (previously 30 seconds). Users can give detailed prompts for specific parts of a song – like intros, verses or choruses – and even control how individual elements should sound.
The model is designed for both creative music and practical use cases like background tracks for videos, podcasts or tutorials. It’s trained on licensed data and won’t mimic specific artists, treating them only as inspiration. All generated content is watermarked with SynthID.
Lyria 3 Pro is rolling out now to paid Gemini subscribers, with daily generation limits depending on the subscription tier and is also expanding to other Google AI platforms.
Source: Android Authority
