AI music creation platform Mureka has launched V9.5, the latest version of its generative music model, with upgrades aimed at producing more natural vocals, coherent song structures and closer alignment with user prompts.
The new model focuses on how different elements of a song, including melodies, harmonies, vocals, instruments and arrangements, work together across a complete track. Mureka said the update is designed to make AI-generated songs sound more structured and expressive rather than simply increasing the complexity of generated audio.
Vocal generation is one of the areas targeted by V9.5. According to the company, the model has been developed to produce smoother phrasing and more expressive vocal performances, while improving how lead vocals interact with other elements of a track.
In internal testing disclosed by Mureka, listeners considered lead vocals generated by V9.5 convincing in 61% of cases. The company also reported that the model followed user prompts in 97% of generations and achieved a 95.7% match rate for requested genres. These figures are based on Mureka's own testing and have not been independently verified.
Beyond vocals, V9.5 is designed to improve arrangement logic and consistency across longer compositions. Instead of treating sections of a song as isolated elements, the model aims to maintain greater continuity between different parts of a track, including changes in dynamics, instrumentation and musical structure.
Mureka said the development of V9.5 was also informed by more than 25,000 responses collected through user surveys. Feedback highlighted areas such as consistency across full-length songs, vocal expression, genre understanding and emotional depth as priorities for further development.
The platform is being used across different stages of music creation, according to Mureka. Some creators use AI-generated material to experiment with melodies and arrangements before moving selected ideas into digital audio workstations such as Ableton for further production. Other users can generate songs from lyrics, concepts or personal ideas.
The release comes as generative AI companies increasingly expand beyond text, images and video into music and audio creation. AI music platforms are competing to improve not only generation quality but also the degree of control creators have over vocals, genres, song structures and production styles.
Mureka V9.5 is now available through the company's music creation platform, while an enhanced version of the model has also been released through its API platform.
Disclaimer: This article may include information derived from interviews, press releases, public statements, research, company communications and other publicly available or third-party sources. Such material may be summarised, paraphrased or contextualised for journalistic and editorial purposes. All rights in third-party content remain with their respective owners.