Inside Suno V6: Custom Models, Voices & My Taste Explained — Build a More Personal Sound
Share
Personalization is not one feature. It is three different layers of creative identity.
My Taste, Voices and Custom Models can all make Suno feel more personal, but they do different jobs. The useful question is not “which personalization feature should I turn on?” It is what part of my creative identity am I trying to make more repeatable?
Repeatable creative identity
Separate taste, voice and sonic system
Rights + model compatibility
Bring the system into Studio
The most important distinction
These tools are easiest to understand when you stop treating them as three versions of the same personalization switch.
My Taste
Preference context.
What kinds of styles, moods and creative directions tend to interest you.
Voice
Vocal identity.
Who is singing — a reusable vocal identity rather than the whole production style.
Custom Model
Sonic system.
A model tuned from your own rights-owned catalog so generations can begin closer to your musical world.
A simple way to remember it
My Taste narrows the neighborhood. Custom Model defines the house. Voice identifies who is singing. Your prompt decides what happens today.
My Taste: teach Suno what you tend to like
My Taste is the broadest personalization layer. Suno describes it as a system that learns from your listening and creation habits — genres, moods and other patterns — and can use that context when you use the Magic Wand in the Styles area. With Style Augmentation enabled, that expansion can reflect what Suno has learned about your preferences.
That makes My Taste especially useful when you know the general direction you enjoy but do not want to write every style detail manually. It can help you get moving faster.
My Taste is currently available to everyone, and Suno provides controls to view, edit or disable it.
Voices: make vocal identity reusable
Voices moves personalization from “what I tend to like” to who the listener hears. Suno introduced Voices as an evolution of its earlier Persona concept, allowing a creator to record or upload their own vocal audio and use that vocal identity in future creations.
Suno also built verification into Voice creation. Its published guidance describes matching the singing voice in the supplied audio with a random phrase spoken by the user. That matters because a reusable voice is not just another style tag — it is identity.
What a Voice should control
Voice should answer: “Who is performing this?”
Voice should not be expected to answer: “What genre is this, what instruments are playing, what is the arrangement, what is the emotional arc, and how should the mix behave?”
That separation is useful. The same vocal identity can move through reggae, soul, trap-gospel, acoustic material or a cinematic arrangement without requiring the rest of the song to remain frozen.
Custom Models: turn your catalog into a reusable sonic starting point
Custom Models are the deepest of these three personalization layers. Suno says Pro and Premier subscribers can fine-tune a model on their own music and then select it from the model picker when creating.
Current Suno guidance says you can build a Custom Model with as few as six songs, create up to three models, use Bulk Upload, and expect model creation to take roughly two to five minutes. Custom Models are private and cannot be shared.
The training set is a creative decision
More songs are not automatically better if the songs contradict each other. If six tracks represent the clearest version of a sound you want to continue, that can be more useful than feeding the model a scattered catalog simply because it exists.
Curate for the identity you want repeated. Think about recurring instrumentation, vocal treatment, tempo range, rhythmic language, density, emotional tone and production character.
Suno explicitly requires that you own the rights to the songs you upload for Custom Model training. Treat that as a hard checkpoint, not fine print.
Which one should you use?
| Your goal | Start here | Why |
|---|---|---|
| I know what I generally like, but not exactly what to prompt | My Taste | Adds preference context to style exploration |
| I want a recurring vocalist | Voice | Keeps vocal identity reusable across different songs |
| I want generations to begin closer to the sound of my catalog | Custom Model | Moves personalization into the underlying sonic starting point |
| I need one specific song and do not need reusable identity yet | Prompt first | Personalization is optional when the task is one-off |
The personalization stack
Once you understand the layers, you can stack them deliberately instead of piling them on.
Preference context → sonic starting point → vocal identity → today’s instruction → deliberate refinement
The order matters conceptually even when the interface does not present it this way. Your Custom Model or Voice does not remove the need for prompting. A repeatable identity still needs a specific song idea, structure, lyric, arrangement direction and editing decisions.
Avoid overpersonalization
Personalization can become another form of over-control. If every generation begins with the same taste profile, the same custom model, the same voice and a highly restrictive prompt, you may eventually remove the very variation that helps you find something new.
A useful pattern is to separate identity work from exploration work. When consistency matters, use the stack. When surprise matters, loosen a layer. Try a baseline V6 generation, experiment with V6-Wild, switch off a personalization layer, or deliberately test an outside reference.
Rights and identity checkpoint
Personalization becomes more powerful when it is connected to real creative assets, which also means the rights questions become more important.
Custom Models: use music you own the rights to. Suno makes this an explicit requirement.
Voices: use your own voice or a voice you are properly authorized to use. Suno’s verification process is a signal that voice identity should be treated seriously.
Compatibility: Suno’s current model guidance warns that Voices, Custom Models and My Taste require compatible models. If a control appears unavailable or behaves differently, check the selected model before assuming the feature is broken.
Private does not mean rights-free. Even when an asset is private inside your account, your source material still needs to be legitimate.
Try this: the 5-minute identity drill
Use one song idea and keep the core prompt as stable as possible. Generate a short test in four passes:
V6 with your prompt only.
Use My Taste/style augmentation.
Use your Custom Model with the same core idea.
Custom Model + Voice with the same core direction.
Do not ask only “which one is best?” Listen for what changed: instrumentation, groove, harmonic language, vocal character, density, phrasing, mood and how quickly the generation moved toward your intended identity. That is how you learn which layer is actually helping.
Part 12 summary
My Taste personalizes preference context. Voice personalizes vocal identity. Custom Models personalize the sonic starting point from a catalog you own. They become more useful when you know exactly which layer you are asking each one to control.
The strongest workflow is not “turn everything on.” It is use the minimum personalization needed to make the identity repeatable, then let the prompt and editing decisions make each song specific.
Suno is one tool. Your creator system has to be bigger than one tool.
We go deep on Suno because AI music creators deserve more than surface-level tips. But JackRighteous.com is built around the larger creator journey: finding your sound, developing your voice, understanding the tools, improving the work, protecting your rights, releasing with more confidence, and building something useful around what you create.
That means the support does not stop when a Suno screen changes. The site connects AI music creation with prompting, genre development, production workflows, creator identity, voice, ownership, creator rights, publishing, branding and the wider AI-creator ecosystem.
If Suno is where you are starting, you can stay here as your skills grow.
Suno Studio 2.0 + V6: move from generation into production
The next step is where these identity decisions meet a fuller production environment: arrangement, recording, clips, effects, MIDI, synths, automation and the Studio workflow.
How do I change models? · What’s new in v6? · Using Custom Models · My Taste · v5.5 Release Notes · Suno v5.5: More Expressive. More You. · Introducing v6
