ElevenLabs just dropped Music v2.0, its next-gen music generation model, with serious upgrades to vocals, instrumentals, and multilingual composition. The big new tricks are inpainting and structured composition. Creators can now select a specific chunk of audio and regenerate it without touching the rest of the track. They can also explicitly define intros, verses, choruses, and outros — a fix for the typical collapse and structural drift that plagues long audio generation.
On the audio quality front, the new algorithm handles dense lyrics and rapid-fire rap, and it can seamlessly switch genres mid-track (think opera straight into heavy metal). The training data is fully licensed, so generated music comes with full commercial rights — no more sync licensing fees or copyright headaches for ads, videos, or brand content.
On the business side, Music v2.0 powers three product lines: ElevenMusic for indie musicians (streaming and remixing), ElevenCreative for brands (downloadable commercial music), and ElevenAPI for developers (high-concurrency generation). And the pricing just got way more attractive: API self-service rates are down 50%, and Creative subscriptions are 40% cheaper.