Kunlun Releases Mureka V9.5: AI Music Model Puts Realism Over Loudness
昆仑万维集团 · wechat · 2026-08-18
Kunlun Wanwei has officially released Mureka V9.5, its new AI music generation model. While competitors race for louder, denser first impressions, V9.5 emphasizes restraint and naturalness—arrangement, vocals and emotion follow realistic music logic, aiming for tracks worth replaying.
Built on the MusiCoT (Music Chain-of-Thought) framework, the model organizes musical "thinking" from full-song structure down to local expression before generating audio. Internal evals show a vocal good-rate of 61.0%, a prompt-control good-rate of 97.0%, and 95.7% full style adherence. Chinese guofeng (national style) tracks improved markedly: clearer lyric articulation, noticeably less "AI flavor" in vocals, and backing harmonies that no longer drown the melody.
The roadmap was shaped by 25,000+ real user feedback submissions. Mureka V9.5 is now available on all platforms worldwide.
Related event: Mureka V9.5 Aims to Make AI Music Sound Human(2 posts)→
More from Multimodal
- Ben Nash's AI-remade Knight Rider starring Elon Musk and Lex Fridman goes viral — bennash · 2026-10-16
- ImageNet should retire: DiT method rankings negatively correlate between ImageNet and text-to-image — YouJiacheng · 2026-10-16
- Open-source anatomy-reveal skill turns Claude into an exploded-view motion design studio — msharmas · 2026-10-16
- Flux 3 Klein vs Flux 2 Klein hands-on: the new base model 'barely usable' for T2I — Both-Rub5248 · 2026-10-16
- USTC paper uses speculative sampling to cut latency in interactive video generation — ustc · 2026-10-16
- 800,000 Gaussian splats run in WebGPU; Casado sees near-infinite scaling — generativist · 2026-10-16