Suno v6 AI Music Models Launch With WMG, BMG, Believe

Suno just dropped v6, its most capable AI music model suite yet, and for the first time the company is calling it a release built "in partnership with the music industry." Warner Music Group, BMG, and French distribution giant Believe are the three partners whose music catalogs fed the v6 training process, alongside behavioral data pulled from how users interact with songs on the platform every day.

The timing matters. Suno is currently valued at $5.4 billion, and v6 is the clearest signal yet that the company is pushing from scrappy AI startup into something that looks a lot more like a formal player in the music ecosystem, lawsuits and all.

What Is Suno v6 and What Can It Actually Do

The v6 rollout includes three distinct models. The standard v6 targets the highest audio quality and most predictable output the platform has ever offered. v6 Mini is a lighter, faster version. Then there is v6 Wild, which CEO Mikey Shulman describes as the model for "peak creativity" over precision. His words: if you want the most interesting sounds you have ever heard and you can live without getting the exact result you asked for, v6 Wild is your tool. If you need something specific, it is probably not.

New features across the suite include editing sections of an existing song using plain language instructions, generating music from text, audio, images, or video, and building mashups from multiple sources in a single request. Suno is also retiring its older models and, as of September 3, has significantly limited downloads of generated music. Both of those moves were first promised in the company's partnership with WMG, announced in late 2024.

The revenue share structure is one of the more interesting details here. Partners receive a share of revenue from v6, but Suno's chief product officer Jack Dolby was careful to say that the payment is not compensation for training data. His framing: the partnership itself is what triggers the deal, not the data contribution. What that distinction means legally, especially given the active litigation against Suno, is a question the industry is still working out.

How the AI Training Actually Works, and Why the Lawsuits Still Matter

Shulman explained the training loop plainly. Suno serves users two generated songs and tracks which one they prefer. That implicit feedback signal teaches the model what makes a song catchier, more compelling, or more listenable. Combine that with the licensed catalog data from WMG, BMG, and Believe, and Suno says v6 is a fully licensed model. Shulman confirmed directly that no Universal, Sony, or Kobalt music trained the v6 models.

That matters because Suno is currently facing multiple active lawsuits. Universal Music Group and Sony Music filed suit in 2024 over copyright infringement. Country artist Jason Isbell is leading a class action claiming Suno designed its generator to mimic artists' identifying characteristics. Mexican label Gerencia 360 Music alleges its regional hits were used without a license. And Round Hill Music is suing over alleged infringement of tens of thousands of songs. None of those cases have been resolved, and the v6 launch does not change that.

Looking ahead, Suno says it is building out AI-enabled remixing, sampling, and other derivative features for music from participating partners. Shulman was direct that no artist's work will be made available for user remixing without consent from both the rights holder and the artist. What those products look like and how the business models will work is still being figured out, though the company confirmed that licensing will be part of the structure when those tools launch.

For independent artists and labels watching this space, the architecture Suno is building matters. The three companies in this first round, WMG, BMG, and Believe, each bring different types of rosters and rights structures to the table. Believe in particular sits closer to the independent and mid-tier artist world than the other two. How those revenue share agreements actually flow down to artists, and whether smaller rights holders eventually get access to the same table, is the thing worth tracking. The remixing and sampling products, when they do arrive, could open up real creative and commercial opportunities for independent catalogs. Or they could be gated behind deal structures that only majors can negotiate.

If you are an independent artist or label thinking about where AI music tools are heading, [check out what's happening in your own city](/events) or [share your work with our community](/submit). The conversation is moving fast and the time to understand it is now, not after the standards are set.