Suno leans into customization with v5.5
The AI music platform's biggest update adds voice cloning and custom model training for personalized tracks.
Suno has released version 5.5 of its AI music generation model, marking a significant shift from previous updates focused on audio fidelity toward user customization and control. The update introduces three core features: 'Voices,' which allows users to train a vocal model on their own voice using acapellas, finished tracks, or direct recordings; 'Custom Models,' enabling training on a personal music catalog of at least six tracks; and 'My Taste,' a system that learns a user's genre and style preferences over time to guide AI outputs.
While 'My Taste' will be available to all Suno users, the powerful 'Voices' and 'Custom Models' features are reserved for Pro and Premier subscribers. The 'Voices' feature, noted as Suno's most requested, includes a verification step requiring a spoken phrase to help prevent voice impersonation, though this safeguard may be vulnerable to existing AI voice cloning tools. Once trained, users can have their AI voice sing on uploaded music or AI-generated instrumentals, while Custom Models let users name and deploy a personalized AI to guide song generation based on their unique sound.
- 'Voices' feature trains AI on user's own voice via uploads or direct recording, requiring a verification phrase.
- 'Custom Models' can be trained on a user's personal music catalog, requiring a minimum of six uploaded tracks.
- 'My Taste' learns user preferences for genres and artists over time to auto-style prompts, available to all users.
Why It Matters
Moves AI music from generic generation to personalized creation, allowing artists to imprint their unique sound and voice directly into the AI.