Suno replaces its AI models with a new one trained on licensed music as copyright suits pile up
Suno introduced Suno v6, a new family of AI music models the company says was trained on licensed catalogs from labels and distributors including Warner Music Group, BMG and Believe. The startup says the v6 models were not trained on the data used for its earlier versions and will replace its older models as it faces ongoing copyright litigation.
Why It Matters
Suno’s shift to models trained on licensed music comes amid multiple lawsuits and recent licensing deals, signaling a potential industry move toward negotiated data use for generative music tools. How labels, artists and platforms respond could shape legal and commercial norms for AI-created music going forward.
Key Facts
- new-model: Suno v6 model family unveiled by Suno
- licensed-data-from: Warner Music Group, BMG and Believe (labels and distributors)
- settlement-with-warner: Suno settled with Warner Music Group last year
- deal-with-bmg: Suno struck a deal with BMG last month
- training-claim: Suno says Suno v6 is not trained using the data used for previous model versions
Suno on Wednesday launched Suno v6, a new set of music-generation models the company says were developed using licensed recordings and related data from labels and distributors including Warner Music Group, BMG and Believe. The startup announced that the v6 family was not trained on the datasets that underpinned its earlier models, and that it plans to retire older versions in favor of the new lineup.
The company is releasing three variants of Suno v6. The base model, positioned for controlled and reliable outputs, will be available to paying customers; Suno v6 wild is an experimental option for ideation and unexpected results and is also sold to paying users; and Suno v6 mini is a faster variant that will be available to all users. Suno says the new models provide greater flexibility in how creators can steer and edit generated tracks.
Suno detailed several new creative features around v6, including the ability to edit parts of a song using prompts or specific lyric words, and to use text, images or video as references when producing tracks. The models can reportedly isolate an instrument from a sample and repurpose it into a new beat. Suno also plans to introduce a remix capability that would only use artists’ songs if those artists opt into a new program the company is setting up with music labels.
The launch comes as Suno continues to face legal challenges from the music industry and others. The company has been sued by labels and artists and still faces litigation from Sony and Universal Music Group as well as plaintiffs such as Jason Isbell and users alleging security lapses. Suno previously settled with Warner Music Group and reached an agreement with BMG. The announcement followed a disclosure the day before that the company had trained some models on YouTube videos. Suno has also added features such as an audio watermark for generated songs and new download limits tied to account tiers, and has raised over $819 million in funding to date, per PitchBook.
Keep Reading
Google DeepMind alumni are building tools to accelerate fusion power for the grid
Besxar is building an orbital semiconductor factory, one SpaceX rocket at a time
Sequoia doubles down on Cymphony as AI agents create new enterprise security risks
