They don’t do short clips anymore.
ElevenLabs just dropped Music v2, a model that actually switches genres in the middle of a song. No glitches. Just smooth transitions from opera to heavy metal, or vice versa if you’re into that chaos.
The model can go from opera to heavy media and back, delivering fast rap without losing its mind.
It’s been ten months since version one launched. That’s a lifetime in AI. Back then? You got snippets. Now, artists can build a whole song by section. Intro, verse, chorus. Stitch them together like a collage. Want to fix the bridge? Isolate it. Re-prompt it. Leave the rest alone.
Fast rap works now too. Coherent. The lyrics stick. Even across languages, the vocals hold up. Add a rain sound effect here. A kick drum there. Non-musical textures mix with the melody.
But let’s be real.
This isn’t just about tech flexes. It’s about legality. Other guys like Suno and Udio are getting sued. Copyright messes. ElevenLabs is banking on licensed data. Commercial use? Clear. They struck deals with labels. That matters more than you think.
Google’s doing its thing at I/O with Flow Music. Stability AI is pushing out complex tracks. Suno’s out there. The race is on for professional-grade generation. Who cares? The output does.
You can edit covers now. Make music videos from prompts. But the real question isn’t who has the best algorithm.
It’s whose tracks won’t get pulled from Spotify next Tuesday.
The music gets weirder every week. Someone will drop a jazz-metal-fusion track by lunchtime. Probably.
Are you ready to hear what AI dreams are about? Or are we still arguing about the ethics while the songs play in the background?
The track is loading.
