Suno releases its first AI music model made with record industry help

AI music startup Suno has launched its latest model, v6, developed in collaboration with major record labels including Warner Music Group, BMG, and Believe.

This updated iteration represents a shift toward using licensed content for training, though the exact composition of the dataset remains somewhat unclear. The release includes three distinct versions: a fast and free version called v6-mini, a standard v6 model, and a variant designed for more unpredictable results named v6-wild. While the technology demonstrates a significantly improved grasp of diverse musical genres compared to its predecessors, it still struggles to replicate deliberate human flaws, such as off-key vocals or disorganized rhythms, and instead tends to produce overly polished outputs alongside typical synthetic audio artifacts.

Beyond performance adjustments, the new generation of models introduces several functional updates designed to streamline the music creation process. Users can now modify specific components of a track, like isolated instruments or lyrics, using conversational text commands without needing to recreate the entire composition from scratch. Additionally, the system enables users to combine elements from multiple existing tracks in their library. Expanding beyond text-only inputs, the latest version also allows creators to generate musical compositions using visual or audio files as prompts. The company plans to phase out older versions as the rollout of the new system continues.