Revolutionizing Music: ElevenLabs' Genre-Switching AI Model
TL;DR
- ElevenLabs has launched Music v2, an updated AI music model that can switch genres mid-track, handle complex vocals, and regenerate only selected sections of a song.
- The system supports section-by-section song building, letting creators compose intros, verses, and choruses separately before stitching them together.
- ElevenLabs says the model is trained on licensed data and is cleared for commercial use, with access through ElevenCreative, ElevenMusic, and soon ElevenAPI.
ElevenLabs’ latest music model is pushing AI-generated audio closer to a real production workflow. Instead of producing only short clips, the company says Music v2 can generate longer compositions, switch styles within a single song, and let users edit or regenerate specific sections without rewriting the whole track.
A new phase for AI music generation
Music v2 is the second version of ElevenLabs’ music-generation system, arriving nearly 10 months after the first release. According to the company, the upgrade is designed to handle both vocal complexity and compositional structure, which are two of the hardest problems in AI music generation.
The headline feature is genre switching. ElevenLabs says the model can move from opera to heavy metal and back again while keeping the song coherent. It can also handle fast rap without losing clarity and can incorporate non-musical sound effects into the track.
How the editing workflow changes
What makes Music v2 especially notable is not just generation, but targeted regeneration. Creators can pick a section of a song and regenerate it with a prompt, while leaving the rest untouched. That makes the tool feel more like a production assistant than a simple prompt-to-song generator.
ElevenLabs also says artists can build songs section by section — for example, generating the intro, verse, and chorus separately, then combining them into a final track. Its documentation for ElevenMusic describes an editing workflow that supports changing lyrics, adding or removing sections, adjusting durations, and applying style keywords for more granular control.
Why this matters for creators
For musicians, marketers, and content teams, this kind of control could shorten the distance between an idea and a usable track. Rather than exporting a full song and starting over when one part feels wrong, users can refine only the portion they want to change.
That matters most for iterative work: ad music, branded content, game soundtracks, social videos, and experimental compositions where mood shifts are part of the design. ElevenLabs is positioning the product for those use cases through ElevenCreative and its new ElevenMusic platform.
Licensed training data and commercial use
ElevenLabs says Music v2 was built using licensed data and is approved for commercial applications. That distinction is important because copyright concerns have been one of the biggest barriers to adoption for AI music tools.
The company is framing this release as a practical option for professional use, not just a demo model. Its documentation says tracks can be exported as high-fidelity MP3 files for professional workflows, and the platform is built for iterative editing and finetuning.
The technology direction
The broader technical trend here is toward controllable generation. Earlier AI music systems often struggled with structure, consistency, and long-range coherence. ElevenLabs claims Music v2 improves consistency across languages, lyrics, vocal styles, and arrangements, suggesting a model that is better at maintaining musical identity over time.
That aligns with the direction of the ElevenMusic product, which is described as an end-to-end workflow for generating songs, refining them, and creating variations from the same prompt. In practical terms, this suggests AI music is moving from one-shot generation toward something closer to a digital audio workstation, but with the model handling much of the composition.
Competitive implications
ElevenLabs is best known for voice AI, so this music push broadens its scope significantly. A model that can handle expressive vocals, stylistic transitions, and editable structure could appeal to creators who want fast turnaround without giving up much control.
If the company’s commercial-use licensing holds up in real-world adoption, Music v2 could become especially attractive to business users who need tracks they can use without clearance headaches. That may make the model less of a novelty and more of a production tool for teams that want speed, flexibility, and legally safer output.
What to watch next
The biggest questions now are about output quality, creative originality, and how well the model performs outside of carefully chosen examples. ElevenLabs says ElevenAPI access is coming soon, which could widen adoption beyond the current product interfaces.
The next phase will likely test whether genre-switching AI music can be useful not just for impressive demos, but for repeatable professional work. If it can, Music v2 may mark a meaningful step toward AI systems that collaborate with creators instead of simply generating from scratch.
Get All The Latest Updates Delivered Straight To Your Inbox For Free!