Suno launches v6, v6-wild, and free v6-mini AI music models built with Warner, BMG, and Believe

Suno released its v6 model family — flagship v6 for Pro and Premier subscribers, a more unpredictable v6-wild for experimentation, and a resource-light v6-mini free to all users — its first models trained with licensed data from Warner Music Group, BMG, and…

On 2026-09-09, Suno launched three music models — v6, v6-wild, and v6-mini — co-developed with Warner Music Group, BMG, and Believe following Warner's November 2025 licensing settlement. Per The Verge, the models were trained from the ground up on a new dataset including licensed content plus user data, though it is unclear whether all dubiously obtained content was excluded. All three variants understand vocals, instrumentation, structure, and mood, accept text, audio, image, or video prompts, and add plain-language editing of individual song sections (lyrics, vocals) and multi-element mashups without regenerating the full track; The Verge's reviewer found v6 still cannot produce intentional imperfections like off-key or dissonant vocals. Universal and Sony continue litigating despite the Warner, BMG, and Believe deals — a Munich court found v3.5 and v4 infringed six works, and the training data size remains under seal in the US District Court in Massachusetts. Suno reports more than 100 million users, over 2 million paying subscribers, and $300 million ARR after raising $400 million at a $5.4 billion valuation in June.

• Three variants launched: v6 (flagship, for Pro and Premier subscribers), v6-wild (more unpredictable, aimed at experimentation), and v6-mini (resource-light, free to all users).
• First Suno models trained with licensed data: built with Warner Music Group, BMG, and Believe following Warner's November 2025 licensing settlement.
• The Verge reports the models were trained from the ground up on a new dataset including licensed content plus user data, but it is unclear if all dubiously obtained content was excluded.
• Models generate songs from text, audio, images, or video inputs and understand vocals, instrumentation, structure, and mood.
• New editing features: text/plain-language editing of individual song sections including lyrics and vocals, and multi-element mashups, without regenerating the full track.
• The Verge notes dramatically improved genre fidelity, but its reviewer found v6 still cannot generate intentional imperfections such as off-key or dissonant vocals.
• Universal and Sony continue litigating despite the Warner, BMG, and Believe deals.
• A Munich court found Suno's v3.5 and v4 infringed six works.