Suno confirms its V6 models were trained in part on 'creations' generated by its own users

Responding to the second copyright lawsuit filed by Sony Music and Universal Music Group in Boston federal court on September 18, Suno confirmed in a September 22, 2026 statement that its V6 model family was trained in part on 'creations' and preference signals generated through interactions with users of its platform, alongside licensed material from its music industry partners. Suno had launched V6 on September 9 as a new generation of models developed with Warner Music Group, BMG and Believe, initially describing its training material as licensed content, user data and accumulated research. The company had previously told Music Business Worldwide that V6's user data did not include audio files uploaded by users - a clarification that did not exclude music generated inside the platform itself. Suno has not publicly detailed the proportion of user-generated music in V6 or identified the individual recordings used. Separately, Suno's terms of service, effective September 3, grant the company a broad, nonexclusive license to use submitted and generated content for purposes including improving its AI models, without automatically transferring ownership of users' original submissions.

Why it matters to creators: anyone who generates music on Suno is, under its terms, contributing material that can train future commercial models - a disclosure that was not part of the V6 launch messaging and that undercuts the 'clean licensed rebuild' narrative.

Not yet known: the precise inputs and proportion of user-generated content in V6; whether training on prior-model outputs 'launders' the alleged infringement from earlier training data, which is the core question of the labels' new lawsuit and has not been decided by any court.