Suno v6 retires earlier models without naming its training catalogue
Suno released its v6 generation in three models, developed with Warner Music Group, BMG and Believe, and plans to retire its earlier models as the service moves to the new generation. The tools can edit song sections, lyrics and audio layers. Yet the company does not identify the recordings used to train v6 or explain how a generated track could be traced to a catalogue.
Artificial Intelligence··Evening
Three models replace the earlier generation
Suno released v6, v6-wild and v6-mini on 9 September. The main v6 model and the experimental v6-wild are available to Pro and Premier subscribers, while the faster v6-mini is open to everyone. The company says it developed the main model with Warner Music Group, BMG and Believe. Suno's announcement and TechCrunch's report confirm the same three partners, the new model family and the planned retirement of the earlier models. Once the transition is complete, the service will run only on the v6 generation, leaving users without the old models for direct comparison inside the product.[1], [2]
Editing tools reach inside the track
The new generation can alter sections of a song through plain-language instructions, make mashups from several sources, sample and isolate audio, build beats and change individual lyrics without rebuilding a track. Text, audio, images and video can also serve as generation inputs. Suno separately describes future opt-in experiences built around individual artists: an artist chooses whether to participate and is paid when they do. The announcement gives no launch date for that arrangement, no payment formula and no method for attributing a finished work.[1]
The partners are named, the training recordings are not
TechCrunch reports that v6 was trained on licensed data from companies including Warner Music Group, BMG and Believe. Suno's own post names those partners but does not list the sound recordings or catalogues used to train v6. Nor does it publish a method for connecting a finished v6 track to a particular catalogue. Retiring the old models ends in-product comparison between model generations, while the announcement's promise of artist participation and payment does not identify the training material or explain how contributions to a finished track would be traced.[1], [2]
Related columns
For more information on this topic, you can read the related columns.
