Suno releases its first AI music model made with record industry help — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Suno's v6 is the company's first AI music model trained with licensed data from Warner Music Group, BMG, and Believe, alongside unspecified "user data" — and it's unclear whether the training set is fully free of dubiously obtained content.
- Suno shipped three v6 variants: v6, v6-wild (designed for "happy accidents and natural imperfections"), and v6-mini (a free tier for fast, resource-light creation).
- v6 dramatically improved genre understanding, nailing superficial hallmarks of hyperpop and krautrock that older models missed by a mile.
- v6 remains incapable of producing off-key, out-of-tune, or dissonant results — prompts requesting "monotone vocals," "no drums," or "fully out of tune" piano were ignored, and vocal artifacts are harsher than in v5.
- v6 adds a chat-based editing interface that lets users modify individual elements like a guitar line or single lyric without regenerating the whole track, plus the ability to mash up outputs from a user's Suno library.
- v6 expands beyond text prompts to accept images, video, or audio as inputs.
- Suno is rolling out v6 today and plans to eventually retire its previous models.
Why it matters: The licensed-data partnership with Warner Music Group, BMG, and Believe signals Suno's attempt to legitimize its training pipeline after earlier copyright disputes, but the inclusion of unnamed "user data" leaves the legal exposure partially unresolved. Free-tier access through v6-mini keeps the model widely available even as Suno courts rights holders.
Ask SkimNews
