On September 9, 2026, Suno released v6, and then it did something that should get every creator’s attention: it announced it is retiring all of its previous models as v6 rolls out. If you learned Suno on v4.5 last winter, got comfortable with v5 in the fall, or finally figured out v5.5 this spring, the specific model behind every preset and prompt you saved is being deprecated. Suno v6 is not an optional upgrade you can ignore. It is the new floor.
That churn is the real story of Suno in 2026. The tool that could “make a full song from a text prompt” back in March still does exactly that, only now the model is better, the licensing is real, the ownership rules are written down, and there is a hard cap on how many finished songs you can download. This is the current, honest picture of what Suno is, how to actually make music with it, and what you are agreeing to when you do.
What v6, v6-wild, and v6-mini actually are
Suno shipped v6 as a family of three models rather than one, and the split matters for how you work.
v6 is the flagship. Suno describes it as reliable and precise, the model that consistently delivers polished output across genres. This is the one you reach for when you want a finished track that sounds intentional.
v6-wild is built for exploration. It is less predictable and more varied, producing textured, ambitious, sometimes strange results. When you are hunting for a sound you cannot describe yet, this is the model that surprises you into it.
v6-mini is the fast, efficient version, and it is the only one of the three that is free for everyone. v6 and v6-wild require a Pro or Premier subscription.
The generation also added capabilities that change the workflow, not just the sound quality. You can now edit part of an existing song using plain language (“change the chorus so it is sung by a gospel choir”) while keeping everything else intact, instead of regenerating the whole track and hoping the good parts survive. You can build a mashup from several songs in a single request. And Suno now accepts multimodal inputs: text, audio, images, and video can all seed a generation. Those three changes, per the official v6 release notes, are the practical reason to care about this version, more than any abstract fidelity bump.
The model you learned last month is going away
Here is the timeline that explains why Suno feels like it never sits still. V3 arrived in March 2024 and was the first version most people tried. V4 followed in November 2024, V4.5 in May 2025, V5 alongside the first Suno Studio beta in September 2025, v5.5 in March 2026, and now the v6 family in September 2026. That is six major model releases in about thirty months, and each one has quietly changed how prompts behave.
I spent 20+ years in IT operations before this, and the pattern is one I have watched play out with every vendor platform: the capability improves, and the thing you built your routine around gets deprecated underneath you. The lesson transfers directly. Do not treat any Suno model as a permanent part of your process. When you make a track you love, export the finished file and the stems while you still can, keep your prompt and lyric text in your own document, and assume the exact model that produced it will not be available next quarter. Suno saying out loud that it will retire previous models is unusually honest, and it is also a warning to build accordingly.
Making a full song, the current workflow
The basic loop has not changed since the “text prompt to full song” days, but each step has more control now.
Start by deciding whether you are writing a Song or using Custom mode. Song mode takes a single description (“upbeat indie folk about leaving a small town, female vocals, acoustic guitar and light drums”) and handles the rest. Custom mode lets you supply your own lyrics, structure tags, and style, which is where you go the moment you care about the result.
Handle lyrics deliberately. You can write your own, let Suno generate them, or paste a draft and use the plain-language editing in v6 to fix a single line without touching the rest of the song. Use structure tags in brackets, [Verse], [Chorus], [Bridge], [Outro], so the model knows the shape you want. This is the single biggest quality lever most creators skip.
Generate, then iterate. Every prompt produces options. Rather than re-rolling from scratch, keep the take that has the right feel and use the section-level editing to repair the parts that miss. This is faster and burns fewer credits than regenerating repeatedly, and it is the workflow the v6 editing tools were designed around.
If you want the track to sound like you, this is where v5.5’s additions come in, and v6 kept them.
Voice cloning and custom style models
v5.5, released March 26, 2026, was the version where Suno stopped being purely a prompt box and started letting you encode your own identity into the output. Three features arrived, per Suno’s own announcement, and they carry into v6:
Voices lets you capture your own singing voice and generate music with it. Suno runs a verification step that matches your singing voice to a spoken phrase, and the resulting voice is private: only you can use it to create new songs. Co-founder and CEO Mikey Shulman framed the whole update around a simple idea, that “the best music starts with a human,” and Voices is the clearest expression of it.
Custom Models let you upload your own tracks to tune the model toward your style. You can build up to three custom models, which solves the “everything sounds like generic Suno” drift that plagued earlier versions when you were chasing a specific genre.
My Taste is the low-key one: the platform learns your preferences over time and personalizes what it surfaces. It is available to all users, where Voices and Custom Models require Pro or Premier.
Suno Studio 2.0 is close to an actual DAW now
If you only know Suno as a website that spits out a finished MP3, Studio 2.0 will surprise you. Shipped August 13, 2026, one day after Suno announced a global licensing alliance with BMG, Studio 2.0 turned the browser tool into something that behaves like a real workstation.
It added MIDI recording and editing directly on the timeline, so you can play notes with a typing keyboard if you do not own a controller, and a built-in wavetable synth for designing your own sounds. Stem separation got better, with new Split from Mix and Advanced Split options you reach by right-clicking a clip. There is a chat assistant that reads your session and edits it on request, and it can even build its own plugins. Studio adds automation curves and seven built-in audio effects. Most important for anyone doing paid work: Premier subscribers export unlimited 32-bit/48kHz multitrack and stems out of Studio, with no download cap.
Studio 2.0 is Premier-only, which makes the top tier a genuinely different product rather than just “more credits.”
What the label deals actually change
For its first two years Suno was, legally, a target. That changed in stages. Warner Music Group settled its copyright lawsuit and signed a first-of-its-kind licensing deal announced November 25, 2025. BMG followed with a global licensing alliance in August 2026, and Believe and TuneCore came aboard around the v6 launch in September. v6 itself was built with Warner, BMG, and Believe, per Digital Music News.
The Warner deal set the template creators should understand. Artists and songwriters keep full control over whether their names, images, likenesses, voices, and compositions are used in AI-generated music. WMG CEO Robert Kyncl put the principle plainly: “AI becomes pro-artist when it adheres to our principles: committing to licensed models, reflecting the value of music on and off platform.” In practice, that opt-in control is why the licensed models are safer to build a business on than the older, litigated ones.
It is not settled, though. Universal Music Group and Sony Music are still plaintiffs, the underlying RIAA case continues, and the American Federation of Musicians has sued UMG and WMG over the settlement terms themselves. If you want the fuller map of who signed what and what it means for the files you make, I broke that down in a separate piece on what the Suno and Udio licensing deals actually give creators.
What you own, and the download math
Two questions decide whether Suno fits your business: can you use the music commercially, and can you actually get the files out.
Commercial rights come with the paid tiers. On Pro or Premier you own what you create and can use it in monetized videos, podcasts, and client work. Free-tier songs are for personal use.
The downloads are where the licensing deals left a mark, and where a lot of creators get caught. As of September 3, 2026, Suno caps how many songs you can download. Free accounts get 7 downloads total, for life, personal use only. Pro gets 20 downloads a month. Premier gets 60 a month, and Premier users exporting from Studio have no limit at all. Credits, separately, do not roll over month to month.
Run that math before you commit. If your channel needs a dozen finished tracks a week, the Pro cap of 20 a month will strangle you, and the answer is Premier plus Studio, not stacking extra download purchases. This is exactly the kind of quiet metered limit I flag in fractional COO work: the sticker price is not the constraint, the throughput cap is.
One more honest point on the YouTube question. AI-generated Suno music can and does play fine in monetized videos when you hold commercial rights, but Content ID is about matching recordings, not detecting that something was AI-made. Keep your download records and your account in good standing, because if you cancel a paid plan, your commercial license for tracks made under it is the thing you need documentation for.
Which tier fits which creator
Free (v6-mini, 7 lifetime downloads, personal use) is for trying it and for hobby projects you will never monetize. Treat it as a demo, not a workflow.
Pro at $10 a month ($8 billed annually), 2,500 credits and 20 downloads a month, is the right tier for most solo creators: YouTubers scoring videos, podcasters making theme music, anyone who needs commercial rights and a steady but modest output.
Premier at $30 a month ($24 annually), 10,000 credits and 60 downloads a month plus unlimited Studio exports, is for people who make music as a core deliverable: producers, high-volume channels, and anyone using Studio’s MIDI and stem tools as a real editing environment.
The company that sits behind those tiers is not a hobby project either. Suno raised a $400 million Series D in June 2026 at a $5.4 billion valuation, more than doubling its $2.45 billion mark from late 2025, and reports over 100 million users with more than 2 million paying subscribers. That funding is worth knowing precisely because it tells you the model churn and the licensing land-grab are going to continue.
Where Suno still is not the answer
For all of that, Suno is not the right tool for every job.
If you need frame-accurate scoring that hits specific cues in a video edit, a tool built for that, like AIVA for cinematic scoring or a stock library, still gives you more precise control than describing timing in a prompt. If you want mood-based background beds with airtight, unambiguous licensing for client work, Soundraw’s copyright-safe approach is simpler to defend to a nervous client. If your project is genuinely collaborative, with other musicians contributing parts, Suno’s single-creator model gets awkward fast. And if you are weighing it against the rest of the field, the honest ranking lives in the 2026 AI music generation roundup, while ElevenLabs for creators covers the voice side that Suno’s Voices feature only partly overlaps.
The practical read on Suno in 2026 is this. The music is genuinely good now, good enough to ship, and the licensing finally gives paid users something solid to stand on. Just go in knowing the model under your feet will change, the download counter is real, and the files you love are worth exporting the day you make them. Build on it like you would build on any vendor platform: use the good parts hard, and keep your own copies.
Recent Posts
Krea 3 Edits a Product Photo in Under Eight Seconds. That Speed Should Change How You Edit.
Krea 3 returns most 2K edits in under 8 seconds at $0.06 an image. The speed matters less as a stopwatch than as a reason to stop writing long prompts and edit one change at a time.
beehiiv's New Pricing Costs Mid-Size Newsletters the Most, and Small Ones Almost Nothing
beehiiv repriced every plan on October 1. Lite at 10,000 subscribers jumped 28% while the Free tier gained podcasts and digital products. Here is the math by list size, the Substack break-even, and...