ASLP Labs debuted DiffRhythm, an open-weights model capable of generating 4-minute songs with vocals

All you have to do is give it lyrics and style reference, and it produces a song—taking just 10 seconds

See it in action:pic.twitter.com/HuaUt7iYK9

— Rowan Cheung (@rowancheung) 5 mars 2025

ASLP Labs debuted DiffRhythm, an open-weights model capable of generating 4-minute songs with vocals All you have to do is give it lyrics and style reference, and it produces a song—taking just 10 seconds See it in action:

→ View original post on X — @rowancheung,