Kunlun Tech has officially announced that on August 18, Mureka rolled out its latest AI music generation model, V9.5. Leveraging the MusiCoT (Music Chain of Thought) framework, this model meticulously constructs a comprehensive musical logic prior to audio generation. It systematically organizes all elements, ranging from the overall composition structure to intricate local expressions. V9.5 places a strong emphasis on arrangement spacing, natural vocal delivery, and emotional restraint, striving for authenticity and engaging listenability in music creation. In internal evaluations, V9.5 has showcased remarkable enhancements in pivotal metrics, including the vocal performance yield rate, control yield rate, and the proportion of complete genre representation. Moreover, the model has refined the precision of lyric articulation and the layering of harmonic arrangements, particularly for Chinese-style compositions. V9.5 is now being progressively introduced across all platforms.