ElevenLabs has released Music v2, a significantly upgraded AI music-generation model capable of shifting genres mid-track — moving from opera to heavy metal and back — while maintaining vocal coherence across fast rap, multilingual lyrics, and complex arrangements.
The model introduces section-by-section composition, allowing artists to build intros, verses, and choruses independently before stitching them together, and to re-generate specific parts using prompts without affecting the rest of the track. Non-musical sound effects can also be embedded directly into compositions.
Crucially, ElevenLabs said the model was trained exclusively on licensed data and cleared for commercial use — a pointed distinction from rivals Suno and Udio, both facing copyright lawsuits from major labels.
The release arrives as Google, Stability AI, and Suno all push into professional-grade AI music generation. Music v2 is available now via ElevenCreative and the ElevenMusic platform, with API access coming soon.
