Suno v5.5 Brings Voices, Custom Models, and Personalized Taste Tools

Suno has pushed AI music generation into personal territory with v5.5, a model update the company describes as "our best and most expressive model yet" and which lets users embed their own singing voice, musical catalog, and listening habits directly into the generation process.

The three-feature release, announced February 6, 2026, centers on Voices, Custom Models, and My Taste. Where earlier updates from Suno focused on improving fidelity and producing more natural vocals, v5.5 shifts the emphasis to user control. CEO and co-founder Mikey Shulman's company framed the release not as a routine technical upgrade but as groundwork for a larger move: Suno's announcement stated the new capabilities "lay the foundation for the next generation of music models we're launching with the music industry later this year."

Voices, available exclusively to Pro and Premier subscribers, is the most distinctive of the three features. Users can record directly into the platform or upload audio of their singing voice, then deploy that voice across Suno-generated tracks. Before the voice becomes usable, Suno requires a verification step in which the user repeats a randomly generated phrase, which the platform uses to confirm voice identity. By default, uploaded voices are private and inaccessible to other users, though Suno has signaled that broader sharing options may arrive without providing a specific timeline. One operational detail warrants attention: if a user publishes a song or shares a direct link with Remix or Cover settings enabled, others gain the ability to use that voice for their own remixes and covers. Voices replaces what the platform previously called Personas, with all prior Personas content migrated automatically to the Voices tab and no previous functionality removed.

Custom Models gives Pro and Premier subscribers the ability to train a personalized version of v5.5 against their own original catalog. Suno describes the intent directly: "you can build a personalized version of the model that knows your style, so the music it makes sounds more like you." Each subscriber can create up to three custom models. Community notes from Suno's platform specify a minimum of six tracks to initiate training, with the model reportedly ready within minutes, though neither figure appears in Suno's official documentation and both should be treated as unconfirmed pending company clarification.

My Taste, the only v5.5 feature open to all users regardless of subscription tier, functions as an algorithmic layer built into the generation interface. Over time, it tracks genres, moods, and sonic references a user gravitates toward, then applies that profile when the user invokes the Magic Wand option in the Styles field. Suno's description of the outcome: "the more you create, the more Suno starts to speak your musical language."

The release lands against a backdrop of intensifying competition. Google DeepMind introduced Lyria 3 Pro, a system aimed at more detailed instrumental rendering and advanced controls for professional creators, making Suno's personalization angle increasingly central to how it distinguishes itself. Suno also disclosed a strategic partnership direction with the broader music industry, having previously announced a collaboration with Warner Music Group in November 2025.

Several questions remain open. Whether the three-model cap on Custom Models is a permanent ceiling is unclear. Suno's language describing it as applying "for now" implies the number could expand. How voice data is stored, how long voiceprints are retained, and what rights users hold over recordings that incorporate an uploaded voice were absent from the v5.5 release materials entirely.

With a user's own voice and catalog now embedded into the generation engine, Suno's claim that v5.5 "doesn't just help create music, but fully reflects the person making it" has shifted from a marketing line into a proposition users can test directly.