Suno has launched its v6 AI music model, marking the first time the company has utilized data licensed from industry partners. The training set includes content from Warner Music Group, BMG, and Believe, alongside user-generated data. The company intends to phase out its previous models in favor of this new release.
The v6 update introduces three distinct versions: the standard v6, v6-wild, and v6-mini. The mini model is available for free and prioritizes speed and resource efficiency. The wild variant is intended to produce less predictable results, though the core v6 model remains focused on maintaining harmonic and rhythmic precision.
New functionality allows users to edit specific song elements, such as guitar lines or lyrics, using natural language commands within a chat interface. This eliminates the need to regenerate entire tracks for minor adjustments. Additionally, users can now combine multiple elements from their existing libraries to create new compositions.
The platform has expanded its input capabilities beyond text prompts. Users can now generate music based on images, video files, or existing audio tracks. While the model demonstrates improved genre recognition compared to its predecessors, it continues to struggle with requests for intentional dissonance or off-key performances.
