Suno v6 is a clean break: licensed-music training, legacy models headed for retirement — the migration window is open

2 sources

Per TechCrunch (2026-09-09T12:05:34+00:00): Suno unveiled the new Suno v6 model family on Wednesday, September 9, 'which it says was developed using licensed data from music labels and distributors such as Warner Music Group, BMG, and Believe', and said v6 'is not trained using the data it used to train previous versions of its music-generating model.' Three tiers: base v6 for paying users (described as reliable and steerable), v6-wild experimental for paying users, and v6-mini — per The Verge, 'the model available for free to all'; legacy models will be retired. Capabilities across both outlets: prompt- or lyric-level editing without regenerating the track, text/image/video/audio references, instrument separation from samples, library-element mashing; v6 started rolling out September 9. Per TechCrunch, remixing is planned — gated on artists opting in to a new program Suno 'is setting up with music labels'; last month Suno said it plans to watermark generated songs. Boundaries: The Verge's hands-on finds genre understanding much improved but 'natural imperfections' still impossible (requests for 'monotone vocals' and 'no drums' get ignored); The Verge also notes the data includes 'user data' and that it's unclear whether v6 training data is 'completely free of dubiously obtained content'; per TechCrunch, the Sony, Universal Music Group, and Jason Isbell suits are still pending. suno.com was unreachable to AITopic on 2026-09-10, so every vendor statement on this card sits at the media layer.

Why now

v6 began rolling out yesterday, September 9, and legacy models are confirmed headed for retirement — creators who build with Suno must make a migration call now: which tier becomes the workhorse, whether back-catalog tracks get redone on v6, whether the free tier (v6-mini) covers their needs. Search interest around a 'tool swap' peaks in the week after release, and licensed-data training is a landmark turn in the AI-music copyright fight — worth explaining precisely while the lawsuits are still open.

Why it is worth publishing

The two sources complement each other cleanly: TechCrunch carries the complete product facts, The Verge supplies ready-to-cite hands-on judgments ('genre hallmarks land; imperfections still don't'). The topic also has native tension: licensed training is the good news, while the still-pending suits and the 'user data' question are the other half of the same sentence — ideal ground for a 'don't just forward the press release' deep-dive.

Evidence basis

Both sources read in full on 2026-09-10: TechCrunch (2026-09-09T12:05:34+00:00, Ivan Mehta) and The Verge hands-on (Sep 9, 2026, 9:42 PM UTC, Terrence O'Brien). suno.com and its blog were unreachable to AITopic on 2026-09-10 (connection timeout), so no Suno first-party page backs this card; all vendor statements stay at the two outlets' layer.

“Suno is replacing its entire AI music stack — the company says the new models are trained on licensed music, and the old ones are going away.”

Angle

Cut it as a migration guide: how to pick a tier (mini as the free entry point, v6 as the workhorse, wild as the lab), which old habits to drop (regenerating whole tracks → editing parts), and the deadline pressure (legacy models will be retired). Layer two is what 'licensed training' actually means: the company says the new data is licensed, but The Verge notes it also includes 'user data' and that it's unclear whether it's fully clean — and the suits haven't stopped. Put 'what changed' and 'what didn't' in the same video.

Format

Short talking-head video

Demo idea

Run three tests on the free v6-mini tier: generate the same prompt on a legacy model and on v6 to compare genre accuracy; edit one lyric line in chat and show the track isn't regenerated; demand imperfection ('monotone vocals', 'no drums') and show whether it gets ignored — the last test doubles as a check on The Verge's hands-on finding.

Platform notes

Attach 'Suno says' or 'per TechCrunch/The Verge' to every training-data statement; never call v6 'fully licensed' or 'lawsuit-proof' — the Sony/UMG/artist suits are still pending (per TechCrunch) and The Verge flags the 'user data' question; remix and watermarking are 'planned,' not shipped; the $819 million funding figure is PitchBook data via TechCrunch — attribute it or drop it; 'its first AI music model made with record industry help' is The Verge's headline phrasing — cite it as such.

Usable claims

Per TechCrunch, Suno on September 9, 2026 unveiled a new model family called Suno v6, 'which it says was developed using licensed data from music labels and distributors such as Warner Music Group, BMG, and Believe', and the company said Suno v6 'is not trained using the data it used to train previous versions of its music-generating model.' Per The Verge, Suno chief product officer Jack Brody said v6 was 'trained from the ground up, with a new set of data that does not include the same data that our previous models were trained on'; that data includes licensed content from Warner Music Group, BMG, and Believe, as well as 'user data' — and The Verge notes 'It's unclear whether that means v6 training data is completely free of dubiously obtained content.'
Per TechCrunch: 'The base Suno v6 model is available to paying users and is said to be reliable and steerable for controlled outputs; Suno v6-wild is an experimental model, also available to paying users, that's meant to be used for ideation and getting unexpected results. Lastly, Suno v6-mini is a faster version available to all users. The company plans to retire its older models.' Per The Verge: 'Mini is the model available for free to all'; prompts are 'no longer limited to text descriptions' because 'Suno v6 can create tracks based on images, video, or other audio'; users 'can edit parts of a song using plain language within a chat, and changing individual elements like a guitar line or a single lyric doesn't require regenerating the entire track'; and v6 'can also mash together multiple elements from your Suno library, combining outputs into new creations.' The Verge's reviewer adds that v6 'nailed all the superficial hallmarks of any genre I threw at it' but 'remains incapable of making anything other than harmonic and rhythmic perfection' — asked for 'monotone vocals' and 'no drums', and the verdict is 'it straight-up ignores you'.
Per TechCrunch, Suno 'plans to add new features such as letting users remix songs — provided the respective artists opt in to a new program it is setting up with music labels that lets Suno use their songs for AI-generated features.' 'I think the music ecosystem and our partners are always looking for ways to create more revenue opportunities for their rights holders and artists. So a big part of this release is creating additional revenue streams there,' Suno chief product officer Jack Brody told TechCrunch. Per the same report: 'Last month, Suno said it plans to add a watermark to songs generated using its platform', and it 'recently introduced new download limits based on the account's tier'. TechCrunch also reports 'Suno is still facing lawsuits from labels like Sony and Universal Music Group, artists like Jason Isbell, and users alleging it ignored security while chasing profits', and that Suno 'has raised over $819 million in funding to date, according to PitchBook data.'

Evidence pipeline

From the news

Breakdown

Suno replaced its entire model family with v6: the company says the new models were trained on licensed data from Warner Music Group, BMG, and Believe, and that legacy models will be retired. This breakdown layers the evidence: the TechCrunch layer (three tiers, part-level editing, the remix and watermark plans, the pending suits, PitchBook's funding figure), The Verge's hands-on layer (Brody's 'trained from the ground up' quote, the 'user data' element, the 'unclear whether... completely free of dubiously obtained content' caveat, better genre handling but still no 'natural imperfections'), and the structural gap on this card — suno.com was unreachable to AITopic, so every vendor statement sits at the media layer. The practical takeaway for creators: the free v6-mini tier is testable today, the migration decision is due now, and nobody on this card says 'licensed means safe.'

Sources

Risks

Say 'Suno says' or attribute to TechCrunch/The Verge for every training-data statement; do not present v6 as 'fully licensed' or 'lawsuit-proof.' When the licensed-data story comes up, name the still-pending suits in the same breath, and link both articles so viewers can read the sourcing split themselves.
Frame remix and watermark as 'planned', not live — demo only what exists in the app today, and note v6-mini is the free entry point. Attribute the funding number to 'PitchBook data via TechCrunch' or drop it, and present the reviewer's quality verdicts as one reviewer's testing.

Demo ideas

Tier-picker card: v6 (paid/steerable) vs v6-wild (paid/experimental) vs v6-mini (free/faster) — advice keyed to what a creator makes, plus the legacy-retirement deadline
'Changed / Unchanged' comparison card: changed — licensed-data training (company-claimed), part-level editing, multimodal references; unchanged — the pending lawsuits, 'natural imperfections' still out of reach (The Verge hands-on), and no clarity that the data is fully clean