> ## Content Index
> Fetch the complete content index at: https://www.metatalks.ai/llms.txt
> Use this file to discover other available public pages before exploring further.

# Suno launches v6 trained on licensed music as copyright lawsuits continue
- URL: https://www.metatalks.ai/suno-launches-v6-ai-music-models-it-says-were-trained-on-licensed-music/
- Published: 2026-09-10T10:17:00.000Z
- Updated: 2026-09-10T10:17:12.000Z
- Author: Al
- Tags: News, AI and Authorship, #newswire

**Suno is retiring its earlier models as it rolls out their replacements. The company says the new generation uses a different training dataset.**

Suno has released v6, a family of AI music models trained on licensed recordings and user data. The launch also puts its licensing deals into operation: chief product officer Jack Brody told Music Ally that Suno would begin sharing revenue with Warner Music Group and BMG when v6 rolled out.

Brody said the training data includes Warner recordings. Suno had begun adding [BMG](https://suno.com/blog/suno-partnership-bmg?ref=metatalks.ai) music, with recordings from participating [Believe](https://suno.com/blog/believe-partnership?ref=metatalks.ai) and TuneCore artists to follow.

The launch comes a day after Suno acknowledged using YouTube videos to train its earlier models. UMG, Sony Music and other rightsholders are still pursuing lawsuits over those models.

Paying subscribers get the flagship v6 and the experimental v6-wild; the faster v6-mini is available to everyone. Brody described a reason for the wild version: tuning models to average user preferences can push their music toward pop. v6-wild is less tuned to those preferences, leaving more room for niche styles.

Suno says v6 lets users change a section or a single lyric through a prompt while preserving the rest of a song. They can also isolate an instrument from a sample and build a beat around it, or use text, audio, images and video as references.

Remixing other artists’ songs will require their consent through an opt-in program with labels. Those tools are still in development.