Voice AI company ElevenLabs has launched a new version of its music-generation model, Music v2, with the ability to switch genres within the same track.
According to TechCrunch, the company said the model was built to handle complex vocals and compositions. The release comes nearly 10 months after the launch of the first version of its music-generation model.
, ElevenLabs said the model can shift from opera to heavy metal and back again, produce rapid rap vocals without losing coherence, and add non-musical sound effects to tracks.
The company added that artists can now select a section of a song and recreate it through prompts without altering the remaining parts of the track.
Instead of producing only short clips, the model allows users to build songs in sections, including introductions, verses and choruses, before combining them into a complete track. ElevenLabs said the model also performs more reliably across different languages, lyrics, vocals and musical arrangements.
In recent months, AI firms have been competing to release models capable of producing professional-grade music. Google, Stability AI and Suno have also introduced music-generation tools capable of creating longer and more complex compositions.
At the Google I/O developer conference, Google introduced features allowing users to create covers, edit songs by sections and generate music videos through its Flow Music tool.
ElevenLabs said the new model was trained on licensed data and cleared for commercial use, allowing users to use generated tracks freely. The company stressed the importance of licensing agreements as AI music firms such as Suno and Udio have faced legal disputes linked to copyright concerns.
The new model is available through ElevenLabs’ ElevenCreative tool for marketing and branding teams, as well as its newly launched ElevenMusic platform for AI-generated songs. Availability on ElevenAPI is expected soon.
.jpg)
