Suno's Creative Sliders: What the Documentation Actually Says
If your Suno style tags seem to be ignored — not weakened, ignored — the most likely explanation is not your prompt. One of Suno's sliders is documented to rewrite your style prompt as its designed function, and it is not off by default. This page goes control by control, quotes what Suno's own documentation says about each, and stops where the documentation stops.
- The Variety slider adjusts and updates your style prompts. That is Suno's own description of it, from the v6 FAQ, and it explains most reports of tags being ignored or expanded without permission.
- Suno states that to retain full control of your style tags, you reduce Variety to 0. That is the single most actionable documented fact about the v6 controls.
- Style Influence is a different control. It sets how closely the output stays to your style input, on a scale from Loose to Strong. It does not rewrite your prompt; it decides how hard the prompt pulls.
- Weirdness runs from Safe to Chaos, and Suno documents 50% as the normal expected result — so the midpoint is the baseline, not a compromise between two extremes.
- A fourth slider, Audio Influence, appears only when you are using an Audio Upload, which is why most users never see it.
- We are not publishing preset numbers. Slider values interact with your specific prompt and genre, and circulated 'magic settings' are the least reliable advice in this category. Change one control at a time against a fixed prompt instead.
Suno style influence, Variety and Weirdness are three separate controls that get discussed as though they were one dial, and the resulting advice contradicts itself constantly. One thread will tell you to lower Style Influence to improve adherence. Another will tell you the opposite. Both are describing real experiences, and neither is describing the control that is most likely causing the problem.
So this page does something deliberately narrow. It takes each control, quotes what Suno's own documentation says about it, and stops there. Where the documentation is silent, we say so rather than filling the gap with a number someone posted.
The one that rewrites your prompt
Start here, because it explains more complaints than the other two combined.
Suno's v6 FAQ describes the Variety slider like this:
"The Variety slider is designed to introduce variety in your outputs by adjusting and updating your style prompts. If you'd like to retain full control of your style tags, reduce the Variety slider to 0."
Read that carefully, because it is doing something people do not expect. Variety does not add randomness to the audio. It modifies the text you wrote before the model acts on it. A non-zero Variety setting means the prompt reaching the model is not the prompt you typed.
That is the mechanism behind the single most common frustration in v6's launch week: carefully constructed style tags that appear to be ignored, or a sparse prompt that comes back expanded into something more elaborate than anything you asked for. The model is not disregarding you. Your input was edited in transit, by a control doing exactly its documented job.
If your tags are being ignored, set Variety to 0 before you rewrite a single word. It is the most actionable documented fact about the v6 controls, and it costs one click to test.
The one that sets adherence
Style Influence is a different control operating at a different stage. Suno's Creative Sliders documentation describes it as letting you choose how close you stay to your style input, on a scale from Loose to Strong.
So the two are sequential rather than overlapping:
- Variety decides what is asked for — it can rewrite the request.
- Style Influence decides how strictly the request is followed — it does not change the words.
That ordering resolves the contradictory advice. Someone who lowered Style Influence and got better adherence was almost certainly fixing something else at the same time, or was describing a different symptom. Lowering a control labelled Loose-to-Strong in order to increase strictness does not follow from the documentation, and there is nothing in Suno's description that supports it.
If your prompt is being followed but too literally — rigid, sterile, obviously assembled from your tag list — Style Influence is the control to move. If your prompt is being changed, it is not.
The one that sets deviation
Weirdness runs from Safe to Chaos, and the documentation includes a detail worth holding onto: 50% is the normal expected result.
That reframes the control. The midpoint is not a compromise between two extremes — it is the baseline, the behaviour you get by default. Moving down makes output more conservative than standard. Moving up makes it less predictable than standard. People who treat 50% as "half weird" and set it to 20% expecting normal output are getting something deliberately more constrained than the default, and often reading the result as flat.
Weirdness and Style Influence are frequently merged into one mental dial, and they are not the same thing. One governs how far from convention the output drifts. The other governs how tightly it holds to your specific input. A track can be conventional and ignore your tags, or unconventional and follow them closely.
The one most people never see
There is a fourth Creative Slider. Suno's documentation notes it in a single line: if you are using an Audio Upload, you also get Audio Influence.
That condition is why most users have never encountered it. It only appears in the audio-upload workflow, where it decides how strongly your uploaded recording steers the generation. If you are building from your own hummed melody or guitar riff, it is the most important control on the page, and our Suno audio input guide covers that workflow in full.
One more piece of context worth having: the Creative Sliders live in Custom mode. If you are working in Simple mode and looking for them, that is why they are not there.
Which control for which symptom
The fastest way to use this page is to start from what is going wrong.
| What you are seeing | The control | What the documentation says |
|---|---|---|
| Style tags ignored or expanded | Variety | Adjusts and updates your style prompts; set to 0 to retain full control of style tags |
| Sparse prompt came back elaborate | Variety | Same mechanism — your text was edited before generation |
| Prompt followed, but rigidly | Style Influence | Sets how close you stay to your style input, Loose to Strong |
| Output drifts from the genre | Style Influence | Move toward Strong |
| Too safe, predictable, generic | Weirdness | Safe to Chaos, with 50% documented as the normal result |
| Incoherent, structure falling apart | Weirdness | Move back toward 50% before anything else |
| Upload barely affecting the result | Audio Influence | Only appears when using an Audio Upload |
| No sliders visible at all | — | Creative Sliders live in Custom mode |
The table is deliberately short on numbers, for the reason in the next section.
Why we are not publishing preset numbers
The most requested thing on this topic is a list of recommended values, and it is the thing we are least willing to give you.
Slider values interact with the prompt they are applied to. A Weirdness setting that produces pleasing unpredictability on an ambient track produces incoherence on a structured pop song. A Style Influence value that holds a detailed instrument map together is unnecessarily rigid on a two-word genre prompt. Circulated "magic settings" are one person's values, for one kind of track, at one moment in a model's life, repeated until they acquire authority they never earned.
There is also a specific error worth naming because it spreads easily: typing a percentage into your style prompt does not set the slider. The sliders are interface controls. Writing "variety 0" into the text box writes the words "variety 0" into your style prompt, which is not the same thing and may well make the output worse.
The honest method is unglamorous and takes about twenty minutes:
- Fix your prompt. Write one prompt and do not change it for the duration of the test.
- Set a baseline. Variety 0, Weirdness at its documented 50% normal, Style Influence wherever you normally leave it. Generate. Keep it.
- Change one control. Move a single slider, regenerate with the identical prompt, and note what changed.
- Record what came back. Where the interface shows you the expanded or adjusted style text, save it — that is direct evidence of what Variety did to your input.
- Return to baseline before testing the next control.
Twenty minutes of that will teach you more about your own material than any preset list, and the result is calibrated to the music you actually make.
Where the settings stop mattering entirely
Every control on this page shapes what the track sounds like. Not one of them affects whether a distributor will accept it, and those are completely separate gates.
Distributor ingestion screening scans for the statistical fingerprint that generation leaves in the audio, plus the provenance data embedded in the file. It does not evaluate the music. v6 was built in partnership with Warner Music Group, BMG and Believe, and Suno has committed to audio watermarking and fingerprinting, so the provenance trail is strengthening. In our benchmark on raw, unprocessed AI tracks: DistroKid rejected 50 of 50, TuneCore 47 of 50 and CD Baby 42 of 50, against approximate thresholds of 0.78, 0.82 and 0.85.
Undetectr works on Suno v6 output and addresses that layer. It processes the audio signal across six layers — spectral artifact removal, temporal pattern normalisation, dynamic range processing, metadata sanitisation, and the removal of embedded SynthID and C2PA watermarks. Because it targets classes of artifact rather than a version-specific signature, a new model generation does not break it, and the watermarking Suno is adding falls inside layers it already covers. Those are the markers DistroKid, Spotify, Apple Music and Amazon Music scan on ingestion.
Tune the settings, clean the file, then distribute. Our guide to uploading Suno tracks to Spotify covers the last part.
How we checked
Every quoted description on this page comes from Suno's own documentation, fetched on 11 September 2026: the v6 FAQ for the Variety slider, and the Creative Sliders help article for Weirdness, Style Influence and Audio Influence. Where we describe a control's behaviour, it is because Suno describes it that way, and where we have stopped short it is because the documentation stops there too.
Community discussion informing which controls confuse people most came from a structured pull of Reddit and YouTube across 12 August to 11 September 2026 — 25 Reddit threads carrying 782 upvotes and 1,351 comments, plus six YouTube videos with transcripts. Since v6 launched on 9 September, direct v6 experience spans two days, and at least one widely shared discussion in that window gives advice on Style Influence that the documentation does not support.
Two caveats. Suno changes interface behaviour without announcement, so re-check the help pages before relying on any quoted description. And we have not run a controlled slider sweep on a fixed corpus ourselves — the method above is what we would do, offered as a method rather than as results.
Questions readers ask.
Because one of the controls is designed to. Suno's v6 FAQ describes the Variety slider as introducing variety in your outputs by adjusting and updating your style prompts, and says that if you want to retain full control of your style tags, you reduce Variety to 0. If you have carefully written style tags and the output ignores or expands them, check Variety before you rewrite anything. This is the most common settings problem mistaken for a prompting problem.
It introduces variation by adjusting and updating the style prompt you wrote, in Suno's own words. It is not a randomness control applied to the audio — it operates on your text before the model acts on it. That distinction matters: a non-zero Variety setting means the prompt reaching the model is not exactly the prompt you typed. Setting it to 0 is documented as the way to retain full control of your style tags.
They act at different stages. Variety modifies your style prompt before generation. Style Influence sets how closely the result adheres to that style input, on a scale Suno labels from Loose to Strong. So Variety changes what is asked for; Style Influence changes how strictly the request is followed. Confusing them is why advice about lowering one to improve adherence to the other circulates and contradicts itself.
It runs from Safe to Chaos, and Suno's documentation states that 50% is the normal expected result. That is worth knowing because the midpoint reads like a compromise and is actually the baseline — moving down makes output more conservative than default, moving up makes it less predictable than default. It is a separate control from Style Influence, which people frequently merge into one mental dial.
A fourth Creative Slider that appears only when you are using an Audio Upload. Suno's Creative Sliders documentation notes it alongside Weirdness and Style Influence with that condition attached, which is why most users never encounter it. If you are working from your own recording, it is the control that decides how much the upload steers the result — our Suno audio input guide covers that workflow.
We are not going to publish preset numbers, because the honest answer is that they depend on your prompt, your genre and what you are trying to fix. Circulated 'magic settings' are the least reliable advice in this category — they are one person's values for one kind of track, repeated until they sound authoritative. The documented facts are worth more than any preset: Variety 0 retains control of your style tags, and Weirdness 50% is the normal baseline. Start there and change one control at a time.
No. Sliders shape what the music sounds like; they have no bearing on distributor screening, which scans for the statistical fingerprint and embedded watermarks that generation leaves in the audio file. Our raw-corpus figures are 50 of 50 rejected by DistroKid, 47 of 50 by TuneCore and 42 of 50 by CD Baby. Undetectr works on Suno v6 output and removes those artifacts so the file clears screening on Spotify, Apple Music and the rest.
The verdict, in one sentence: Undetectr.
Sliders decide what the track sounds like. They decide nothing about whether a distributor accepts it. Undetectr works on Suno v6 and removes the AI watermarks and spectral fingerprints that DistroKid, Spotify and Apple Music scan for. $39 one-time, roughly 90 seconds per track.