AI music startup Suno released version 5.5 of its music generation model on March 26, introducing three features that shift the tool from generic generation toward personal expression.

Voices is the headline feature: users can now record their own voice and have AI-generated songs sung back in that voice. Suno says the model captures the character of your voice rather than cloning it note-for-note, letting you perform music you couldn't otherwise play.

Custom Models lets users fine-tune the base model on their own existing tracks. The idea is to produce output that reflects a specific style or catalog rather than Suno's default sound — a feature that could appeal to producers who want AI assistance without losing their sonic identity.

My Taste is a passive personalization layer. As you use the platform, the model learns which genres, tempos, and moods you gravitate toward, and adjusts its defaults accordingly. It's a small change but signals a longer-term ambition: an AI music collaborator that gets better the more you use it.

The v5.5 announcement follows a period of increasing competition in AI music generation. Google DeepMind's Lyria 3 Pro, released earlier this week, supports structured long-form tracks up to three minutes. Suno's response focuses less on raw capability and more on making output feel distinctly personal.

The update is live for Suno subscribers. Custom Models and Voices are rolling out in phases.