Suno replaces its AI music model lineup with v6, trained on licensed label music

Suno v6 swaps unlicensed training data for label deals

AI music startup Suno has launched a new model family called Suno v6, and according to TechCrunch, the company says it was developed using licensed data from labels and distributors including Warner Music Group, BMG and Believe. Suno also states that v6 was not trained on the data used for earlier versions of its music generation models, a notable shift for a company that has spent the past years fighting accusations of training on copyrighted recordings.

The launch follows a period of legal and commercial maneuvering. As TechCrunch reports, Suno settled with Warner Music Group last year and reached a deal with BMG last month, though lawsuits from other parties are still active.

Three tiers, and old models are on the way out

Suno is shipping v6 in three variants. The base Suno v6 model is available to paying subscribers and is positioned as the dependable option when users need controlled, predictable output. Suno v6 wild, also limited to paying users, is an experimental version aimed at ideation and less predictable creative results. Suno v6 mini is a faster cut available to everyone, including free accounts. The company intends to retire its older models, meaning the licensed v6 family becomes the core of the product.

The new lineup also expands what users can do with generated tracks. Suno v6 supports editing a specific section of a song using a prompt or even a single word from the lyrics, and it can take text, images or video as reference material when creating music. Users can also isolate an instrument from a sample and build a new beat around it.

Remixes, but only where artists opt in

Suno said it plans to introduce a remixing capability, but with a catch: artists will have to opt in through a program Suno is setting up with music labels before their songs can be used in AI-generated features.

Jack Brody, Suno's chief product officer, told TechCrunch that partners in the music ecosystem are consistently looking for new ways to generate revenue for rights holders and artists, and that a major goal of this release is opening up additional revenue streams. He also framed derivative work as a revenue opportunity for all stakeholders.

Beyond licensing, Suno has been tightening other parts of its platform. TechCrunch notes the company said last month that it plans to watermark songs generated on its service, and it recently introduced download limits tied to account tiers. Brody suggested Suno can help address ecosystem problems such as streaming fraud and mass exporting to low-intent distributors, while ultimately leaving content governance to the platforms themselves.

Legal exposure has not gone away

The licensed pivot does not resolve Suno's legal situation. The company still faces lawsuits from Sony and Universal Music Group, from artists including Jason Isbell, and from users who allege the company neglected security while pursuing growth. TechCrunch also points out that the v6 announcement landed just one day after Suno admitted that it had trained its models on YouTube videos.

Despite the controversies, Suno has raised more than $819 million to date, according to PitchBook data cited by TechCrunch.

Why it matters

Suno's move from disputed training data to licensed label catalogues is one of the clearest examples yet of a generative AI company rebuilding its core product around rights-holder deals rather than fighting in court indefinitely. If v6 performs well commercially, it could become a template for how AI music, and possibly generative media more broadly, gets legitimized through revenue-sharing arrangements with incumbent rights owners.

But the unresolved suits from Sony and Universal, plus the fresh YouTube disclosure, show the transition is incomplete. Suno is betting that licensed training data and opt-in artist programs will satisfy the market before the courts decide whether its earlier practices were lawful.