Suno Releases v6 Audio Models with Stem Separation and In-Track Lyric Editing
Suno has launched its v6 generative music architecture, developed using licensed data from music labels and distributors including Warner Music Group, BMG, and Believe. The release adds granular lyric-level audio editing, stem separation, and multimodal inputs.
Impact: Medium
Why it matters
As generative music services face copyright litigation from major labels, Suno v6 reflects a transition toward commercial licensing partnerships alongside more granular creative controls like stem separation and audio inpainting.
TL;DR
- 01Suno v6 was developed using licensed data from partners including Warner Music Group, BMG, and Believe.
- 02The model family splits into Base (steerable), Wild (experimental ideation), and Mini (high speed) variants.
- 03Creators can isolate instrument stems from samples and edit audio sections via lyrics or prompts.
- 04To mitigate distribution spam, Suno enacted tier-based download limits and announced plans to add watermarking.
Key facts
- Model Variants
- Suno v6, Suno v6 wild, Suno v6 mini
- Licensing Partners
- Warner Music Group, BMG, Believe
- Input Modalities
- Text prompts, lyrics, images, video references
- Total Disclosed Funding
- $819M (PitchBook)
Model Lineup and Access Tiers
Suno v6 marks a departure from the training data used for its previous versions, which the company plans to retire. The v6 release splits into three distinct versions:
Suno v6(Base): Targeted at paying subscribers, prioritizing steerability and reliable, controlled outputs.Suno v6 wild: An experimental release for paying users aimed at ideation and unexpected musical results.Suno v6 mini: A faster version available across all user tiers, including free accounts.
Granular Audio Editing and Stem Extraction
Rather than forcing creators to regenerate entire tracks, v6 introduces localized editing. Users can edit a part of a song using a prompt or by selecting a word in the generated lyrics.
Additionally, v6 lets users separate an instrument from an audio sample and reuse it to build a new beat or foundational track.
Multimodal Prompting and Distribution Safeguards
The architecture now accepts text, images, or video files as references to shape track generation. To help manage broader ecosystem issues such as streaming fraud and low-intent bulk uploads, Suno recently introduced download limits tied to account tiers and stated that it plans to add audio watermarking to generated songs.
✓ When to use
- Rapid song prototyping and beat-making using extracted stems.
- Soundtrack ideation using video or image mood boards as references.
✕ When NOT to use
- When complete commercial indemnity and clear copyright status are strictly required for corporate publishing.
- When offline or local audio generation is needed.
What to do today
- Test lyric-level editing in Suno v6 to adjust specific song sections without full track regeneration.
- Experiment with multimodal inputs by providing video or image references for thematic arrangement.
Sources