- Suno’s v6 family includes v6, v6-wild and v6-mini, with previous models scheduled for retirement. <sup>[1]</sup><sup>[2]</sup>
- Suno says v6 was developed with Warner Music Group, BMG and Believe and was not trained on the data used for its earlier models. <sup>[1]</sup><sup>[3]</sup>
- The company has described opt-in artist products and partner compensation, but has not published a complete v6 training-data list or a payment formula for individual artists. <sup>[1]</sup><sup>[4]</sup>
- Copyright and labor disputes involving Suno’s earlier systems remain active, including claims from independent artists, SOCAN and the American Federation of Musicians. <sup>[5]</sup><sup>[6]</sup><sup>[7]</sup>
- Stable Audio offers a more detailed public comparison by identifying licensed or Creative Commons training sources and publishing output terms. <sup>[8]</sup><sup>[9]</sup>
Suno has launched a new generation of AI music models that it says was developed with licensed music from Warner Music Group, BMG and Believe, as the company moves to retire the systems at the center of multiple copyright disputes. [1][2][3]
The company says v6 was not trained on the data used for previous versions of its music generator. That distinction could matter in litigation, but the launch does not itself resolve claims about how older models were trained or what they generated. [3][5]
What is changing #
Suno’s v6 family includes v6, its flagship model for Pro and Premier subscribers; v6-wild, a more experimental paid model; and v6-mini, a faster version available to all users. The models can edit sections of a song, change individual lyrics, combine user-provided musical elements and use text, audio, images or video as inputs. [1][2]
Suno says it will retire previous models as v6 rolls out. An August policy update said songs already created would remain in users’ libraries and continue to be playable and shareable, including for covers and remixes. [2]
The company did not provide a date-by-date retirement schedule in its launch announcement. It had previously said that all prior models would be retired when the new generation launched, so the transition is underway but its completion date remains unspecified. [2]
Licensed, but not fully itemized #
Suno’s announcement names Warner Music Group, BMG and Believe as industry partners. Its BMG announcement said the deal covers BMG’s recorded and publishing repertoire and that participating artists and songwriters would have their rights protected and be compensated. Warner’s earlier partnership announcement said artists and songwriters would control whether and how their names, images, likenesses, voices and compositions were used in new AI-generated music. [1][4][10]
Suno has not published a complete list of recordings or compositions used for v6, the share of the training corpus supplied by each partner, or the payment terms for artists, songwriters, performers and other contributors. Its launch post instead describes future opt-in products built around individual artists, saying participants will be paid when they choose to take part. [1]
That leaves a distinction between company-level licensing and individual creator compensation. A label or distributor may have authority to license catalog material, while performers, session musicians and songwriters may have separate contractual or labor claims. The American Federation of Musicians is pursuing that issue in a lawsuit against Universal Music Group, Warner Records and Atlantic. [6]
The legal cases are not erased #
In Justice v. Suno, a Massachusetts federal judge allowed copyright and Digital Millennium Copyright Act claims by independent artists to proceed past the motion-to-dismiss stage. The plaintiffs alleged that Suno downloaded copyrighted songs, used them to train its model and generated music that was identical or substantially similar to protected works. The court dismissed a Tennessee consumer-protection claim. [5]
SOCAN, Canada’s performing-rights organization, filed a separate lawsuit on September 2, 2026, alleging that Suno generated and publicly streamed outputs that replicated works in SOCAN’s repertoire without consent or compensation. [7]
The new licensing arrangements may affect future model development or claims involving partner catalogs, but that is not the same as resolving allegations involving earlier models or works outside those agreements. The legal effect will depend on the terms of each settlement or license and on the claims in each case. This is an editorial inference from the separate rollout and litigation records. [3][5][6][7]
Safeguards and the rival comparison #
Suno says it has added safeguards to screen uploaded audio and lyrics for unauthorized use and make abuse harder. Its chief product officer separately told Axios that the company continues to block direct prompts for a song in the likeness of a known artist or existing song. Suno’s updated terms also limit downloads, including 20 monthly downloads for Pro users and 60 for Premier users, while saying paid subscribers retain commercial rights to downloaded songs. [1][2][11]
Stable Audio provides a useful comparison because Stability AI has publicly identified parts of its training approach. The company says Stable Audio 2.0 used more than 800,000 AudioSparx files under license, while Stable Audio 3.0 is described as trained on fully licensed data and offers users output rights under its community or enterprise licenses. [8][9]
Suno has disclosed its commercial partners and broad opt-in plans, but not an equivalent public accounting of the v6 corpus or a detailed royalty framework. Until those details are published, “licensed” describes the company’s stated development framework more clearly than it describes the economic path from the model to every creator whose work may be involved.
Further sources #
The stories that matter, in one email. Free — unsubscribe anytime.