On May 27, 2026, ElevenLabs released Music v2, its second-generation music model, roughly ten months after the original arrived on August 5, 2025. The headline capability is genre switching: a single track can move “from opera to heavy metal and back” without breaking musical coherence. Just as significant is the licensing posture — the model is built on licensed data and explicitly cleared for commercial use, a deliberate line drawn between ElevenLabs and the lawsuit-mired Suno and Udio.
For content and marketing teams, this moves AI music from novelty to usable stock: no sync fees, no clearance waits, just background scores for ads and branded video.
A Second Generation, Ten Months Later
The first model produced short clips, which capped its usefulness. Music v2 adds full-song construction: you build section by section — intro, verse, chorus — and the model stitches everything together while keeping structure and continuity intact, instead of splicing unrelated samples. ElevenLabs also claims across-the-board improvements in vocals, instrumentation, arrangement, and multilingual support, and says the model can sustain rapid-fire rap delivery without losing coherence.
When the original model launched in August 2025, ElevenLabs’ pitch was already commercial clearance rather than raw fidelity. Ten months later that pitch has widened from “safe to use” to “safe, long-form, and steerable” — a fair summary of how fast production requirements moved.
Genre Switching Within a Single Track
Genre switching is the marquee feature. Mid-track, you change the style and the model makes the transition sound natural rather than like two songs glued together. For film and game scoring this is immediately practical: the music follows a scene change instead of an editor manually cutting between clips.
One more detail that will matter to producers: non-musical sound effects can be embedded directly in the generated track, so ambient or impact audio no longer requires a separate overdub pass.
Section-by-Section Builds and Inpainting
Music v2’s inpainting lets you select any section and regenerate only that portion from a prompt — rewrite the bridge, keep the chorus untouched. That turns music generation from a one-shot dice roll into an iterative editing workflow, closer in spirit to video editing or photo retouching than to slot-machine generation.
For teams producing volume content, iteration matters more than raw quality. A model you can steer section by section compounds: ten controllable generations beat a hundred uncontrolled ones.
Licensed Data, Cleared for Commercial Use
ElevenLabs says Music v2 was developed “with artists, labels, and publishers,” building on rights-respecting licensing collaborations including Believe. That positioning lands with force in 2026: Universal and Sony are seeking to add roughly 61,000 copyrighted works to their lawsuit against Suno after discovery showed training on millions of their recordings, and Sony’s suit against Udio has added more than 30,000 recordings.
While rivals burn cycles in court, ElevenLabs is selling certainty: a generated track you can ship in a client campaign without wondering whether a takedown notice arrives next quarter. For agencies and brands, that certainty is the product.
The Competitive Field and Price Cuts
The AI music market heated up fast in the first half of 2026. Google’s Lyria 3 Pro launched March 25. Stability AI shipped an audio model capable of six-minute songs on May 20. Suno is on v5.5, and Flow Music at Google I/O added covers, section editing, and music videos. Accelerating model cadence is the defining pattern of this year — a thread we flagged in our opening outlook for 2026.
ElevenLabs answered on two fronts: capability and price. ElevenAPI pricing drops by up to 50%, and ElevenCreative self-serve plans fall by up to 40%. Music v2 is available now in ElevenMusic — the consumer app launched April 2, 2026 — and in ElevenCreative for marketing and branding work; the API is listed as “coming soon,” with early access through the sales team. The split matters: ElevenMusic covers listening, remixing, and creating, while ElevenCreative targets advertising and branded content, the segment where sync licensing costs have historically bitten hardest.
For developers and content teams, the shortlist for AI music tooling this quarter will hinge less on which model sounds best and more on two harder tests: licensing certainty you can defend to a legal department, and API availability you can actually build against.
Sources
- ElevenLabs’ new music-generation model can switch genres mid-track — TechCrunch
- Introducing Music v2 — ElevenLabs
- ElevenLabs releases a new AI-powered music generation app — TechCrunch
AI-assisted summary compiled from the sources above, reviewed by a human before publishing.