Leading AI music generation platform Suno recently released its milestone v5.5 version, the most significant update in its model history. Unlike previous iterations that focused on audio quality and natural vocal performance, this update centers on fully empowering users with control, introducing three groundbreaking features: "Voices," "Custom Models," and "My Taste," marking a shift in AI music creation from standardized output to deep personalization.

In the feature matrix, "Voices" is officially defined as the most anticipated feature. This feature allows users to train a personalized acoustic model by uploading a cappella audio, instrumental tracks with background music, or live recordings. To balance convenience and security, Suno specifies that higher-quality recordings require less data, and it has introduced a verification phrase mechanism to prevent voice misuse. Once training is complete, the AI can simulate the user's vocal style to sing any generated or uploaded song. Additionally, the "Custom Models" feature for professional users requires at least six original tracks to enable the model to accurately capture and learn the user's creative style.