Suno Launches v6 Music Model Family Built With Warner, BMG, and Believe, Retires All Prior Versions

Suno has released its v6 generation of AI music models, built with label partners Warner Music Group, BMG, and Believe, and is retiring all previous model versions from the platform, the company announced in a blog post from CEO and cofounder Mikey Shulman.

The rollout comes in three variants. The flagship v6 targets paying Pro and Premier subscribers, promising, according to Suno, reliable and precisely polished output across genres. v6-wild, also restricted to subscribers, is described as an experimental version that produces less predictable, texture-rich results meant as a starting point for further editing. v6-mini is free for all users and is positioned by Suno as faster and more efficient than any other free model on a music platform.

All three models are multimodal, generating songs from text, audio, image, or video input. Suno says the models now understand more musical building blocks than prior generations, including vocals, instrumentation, song structure, mood, and stylistic references.

Text-based editing and mashups

v6 allows users to edit individual parts of an existing song through text commands without regenerating the full track — for example, replacing a chorus vocal with a gospel choir or altering individual lyric lines. Users can also combine elements from multiple sources in a single request, such as pairing vocals from one track with drums from another and new lyrics styled after '80s synthwave. The model can process voice recordings, video, and images into a single song from one prompt, and can isolate a sample to build a new beat around it.

This update replaces v5, released in September 2025, and the interim v5.5 voice-cloning update from March.

What Suno won't disclose

Suno calls Warner, BMG, and Believe "codevelopers" but has not disclosed which catalogs were used or how much material went into training v6. The company had asked the U.S. District Court in Massachusetts to keep the size of its training data sealed, arguing competitors could infer its development approach from the figure.

Safeguards cited for the launch — first announced in August — include removing artist names from training metadata, screening uploaded audio and lyrics for unauthorized use, and a policy against mass distribution of generated songs on streaming platforms. Suno says songs made with the partner-backed model will be distributable through Believe and TuneCore. The company also plans future offerings built around individual artists, who would opt in and be compensated according to Suno.

Warner settles, Universal and Sony keep suing

The partnerships stem from Suno's November 2025 settlement with Warner Music, under which Suno committed to building licensed successor models. BMG and Believe followed with similar deals in August and early September 2026. Universal Music Group and Sony Music remain plaintiffs in the case coordinated by the RIAA in June 2024 and have not settled.

Suno raised $400 million in June at a $5.4 billion valuation. In late July, the Munich Regional Court I found Suno liable for copyright infringement, ruling that six musical works were reproducibly embedded in its v3.5 and v4 models. Suno claims more than 100 million people have used the platform to create songs, with over 2 million paying subscribers and annual recurring revenue that topped $300 million in February.

What this means

Suno's licensing deals with Warner, BMG, and Believe mark a shift from the industry-wide legal confrontation that defined generative music's first wave toward negotiated coexistence — at least with some labels. But Universal and Sony's continued litigation, combined with Suno's refusal to disclose training data details and the Munich court's infringement finding against earlier versions, means the underlying copyright question for v6 remains unresolved rather than settled. Whether licensed inputs meaningfully change what the model can legally reproduce is a claim only a court, not a blog post, can verify.