Suno v5.5: Personal Voice, Taste Filters, and Custom Models

Suno released v5.5 — the largest update to its AI music generation model since its launch. While previous versions focused on sound quality and vocal naturalness, v5.

5 is entirely devoted to user control over the result. The update includes three fundamentally new tools: Voices, My Taste, and Custom Models. Voices is the most anticipated feature according to the company itself.

It allows you to train a vocal model on your own voice: users can upload a clean a cappella, a ready track with instrumentals, or simply sing directly into the microphone of a phone or laptop. Training quality depends directly on recording clarity — the better the material, the less data required. To protect against abuse, Suno has built in a verification mechanism: the system confirms that the voice belongs to the user uploading it.

My Taste is a personal taste filter. The user specifies their music preferences, and the model shifts generation toward familiar styles and genres. This solves one of the key problems with AI music: the result often sounds "averaged" due to the broad scope of training data.

Custom Models go even further: users can create their own specialized models for a specific genre, style, or sound aesthetic. This opens up possibilities for musicians, brand directors, and content developers who need a recognizable branded sound. Suno continues to grow its user base amid mounting pressure from the music industry.

Several major labels have sued the company, accusing it of unauthorized use of protected recordings for model training. Despite this, Suno continues to aggressively develop its product. V5.

5 changes the positioning logic: from a tool "press and get music" to a platform of personalized AI creativity. The ability to sing in your own voice and customize the model to your own taste is a step toward professional users, not just those who want to quickly create a background track.