Moonshot AI's Kimi K3 Triggers Selloff in Chinese Rivals
Beijing-based Moonshot AI has released a record-breaking model that outperforms mid-tier US systems on key tests, wiping out market value from competing Chinese AI firms.
Moonshot AI on Friday unveiled Kimi K3, a 2.8 trillion parameter model that the Beijing-based startup claims outperforms Anthropic’s Claude Opus 4.8 and OpenAI’s GPT 5.5 on coding and general agent benchmarks. The system still trails the absolute leading-edge US offerings, specifically Anthropic's Claude Fable 5 and OpenAI's GPT 5.6 Sol, in overall performance. However, it stands as China’s largest AI model to date.
The release immediately punished the stock prices of competing Chinese AI developers. Z.ai, which released a widely publicized model in June, saw its shares plummet 28% on Friday. MiniMax Group fell 16%, while Alibaba dropped 4%, erasing gains made earlier in the week on news of its Apple partnership.
The market reaction underscores a rapid reshuffling of the competitive landscape among Chinese artificial intelligence labs. "K3 raises the capability ceiling for China AI models, shifting the burden of proof to other independent AI labs," said Bank of America analyst Alex Liu in a note to clients. The selloff suggests investors are reassessing the moats of companies that previously held the top tier of the domestic market.
For established tech giants, the debut presents a specific strategic challenge. "For Alibaba, while it benefits from broad AI training/usage growth for its cloud service given tight compute environment, Alibaba Qwen's 'open-source leader' narrative may face some tests," Liu noted. Alibaba and Tencent are both backers of Moonshot AI, which raised $2 billion at a valuation exceeding $20 billion in May.
The model's arrival highlights the intensifying race for artificial intelligence supremacy between the US and China. Chinese models are steadily closing the performance gap with American systems while retaining a significant cost advantage. This dynamic is driving adoption among Western companies, a trend that has prompted US lawmakers to consider curbing the use of Chinese AI domestically.
Bank of America's analysts emphasized that the breakthrough was achieved despite severe local limitations. "Despite persistent hardware/compute capacity constraints in China, K3 demonstrates that pre-training scaling, paired with architectural innovation, can still deliver step-change gains for flagship Chinese models," Liu wrote.
