ElevenLabs BlogJun 8, 2026, 9:02 AMimportant 72

Introducing Music v2

ElevenLabs introduced Music v2, a new AI music model with stronger vocals, arrangement, multilingual support, and editing control.

ElevenLabs says Music v2 improves vocals, instrumentation, arrangement, and multilingual generation across genres. The model adds finer control through inpainting, section-by-section song building, reference matching, and more complex vocal or stylistic handling. It powers ElevenMusic and ElevenCreative now, with ElevenAPI access coming soon, and is positioned for musicians, developers, brands, and content teams.

ElevenLabs introduced Music v2 on its blog, a next-generation AI music generation model focused on music quality, controllability, and commercial use cases. According to the company, Music v2 improves on its predecessor in vocals, instrumental performance, arrangement, and song structure, and it also strengthens multilingual support, so that when users write lyrics in different languages, the model can more reliably generate vocals, lyrical delivery, and overall arrangement in the corresponding language. On the feature side, Music v2 doesn't just generate short clips—it can build complete songs section by section, such as intro, verse, chorus, and so on, while maintaining structure and continuity. It also supports finer editing: users can select a specific section of a track and regenerate it, for example modifying only the bridge while keeping the chorus and other parts intact. ElevenLabs also emphasizes that the model can handle more complex musical demands, including spanning styles such as opera and heavy metal within a single song, maintaining fast rap or dense lyrics, and even embedding non-musical sound effects into the music without breaking overall musical coherence. In terms of product positioning, Music v2 will power three ElevenLabs platforms: ElevenMusic, aimed at musicians and creators, which lets them create and remix starting from lyrics, mood, or a reference track; ElevenAPI, aimed at developers, which lets them integrate music generation, inpainting, reference matching, and other capabilities into products; and ElevenCreative, aimed at brands and content teams, used for ads, branded content, and video music. The company also states that Music v2 is already available in ElevenMusic and ElevenCreative, while ElevenAPI is coming soon or offered via early access; at the same time, prices for both Music v1 and Music v2 on ElevenAPI and ElevenCreative have been reduced. Notably, ElevenLabs claims the model was trained only on licensed data, that generated tracks can be used commercially, and it emphasizes no sync fees and no clearance delays. For Taiwanese readers, the key point of this article is not research details but the fact that ElevenLabs is pushing AI music from one-off generation toward an editable, embeddable, commercially-licensed workflow.

Full summary

Free shows the 3-line summary; Pro unlocks the full deep summary (~300 words) so you never have to click through.

See Pro plans →

Want the original English / full article?

Read on ElevenLabs Blog →

Summaries are AI-generated; the original article is authoritative.