No more anonymous artist: enter "Voices" cloning
Until now, when you asked an AI to create a song for you, it generated a random male or female voice. The explosive Suno v5.5 update (rolled out in late March 2026) completely changes the ecosystem with its new Voices feature.
The principle is worthy of a science-fiction film: you simply sing a few notes (without music) for a few seconds into your computer's microphone. Suno then mathematically "captures" the unique timbre of your voice and its emotion. Then what? You type in the lyrics of a song, and it is literally your own human voice that will start singing with the technical perfection of a great studio artist, whether on a hyper-fast rap flow, pop or extreme Death Metal!
Custom Models
Only yesterday, generating a song on Suno was a bit like playing the lottery: you typed "sad guitar music", and you crossed your fingers. Today, the software adapts to you, like a tailor taking your measurements.
You can upload a few of your own existing creations. The AI will then listen to them carefully, "dissect and learn" your personal style (how your drums sound, how you place your silences), and create an AI entirely cloned from your own musical DNA. So the AI no longer makes a sad guitar song, it reliably makes your sad guitar song.
The "My Taste" feature: the AI that reads your mind
Just like the recommendation algorithm that instantly pins down your video tastes, the My Taste feature learns what you particularly like to create. If the intelligence notices that you often generate 1980s electronic music with big synthesizers, your future creations will organically take that aesthetic direction, without you even needing to specify it explicitly in text.
Faced with the entirely justified anger of the music industry over the theft of famous artists' voices, Suno has heavily reinforced its security. You can clone only your own voice via a strict live verification through your phone's or PC's microphone. In 2026, it is technically and legally impossible to import an old MP3 file of Ed Sheeran or a well-known singer to force them to sing.
How much does it cost? (Pricing)
To take advantage of the voice cloning feature and the custom models (which are the real big massive new features of v5.5), you need to move to a paid plan to cover the heavy computing costs of the servers. Here is the summary for creators:
| Plan | Price | What the plan includes |
|---|---|---|
| Basic | Free | 50 credits/day (basic creation, no cloning, personal use only) |
| Pro | $8 or $10 / month | 2,500 credits. Access to Voice Cloning (Voices) and Custom Models, with full Commercial Rights. |
| Premier | $24 or $30 / month | 10,000 credits, priority and faster generations. |