China’s top AI developers have fired off a rapid sequence of model releases that challenge Silicon Valley’s hold on frontier artificial intelligence. Within days, Beijing-based Moonshot AI and tech giant Alibaba unveiled systems they claim can match or beat the best from OpenAI and Anthropic — at a fraction of the cost. The moves signal that America’s edge in AI is shrinking fast, just as the technology becomes central to national security and economic competition.

Moonshot dropped Kimi K3 on Friday, saying internal tests rank it above nearly every U.S. system except OpenAI’s GPT-5.6 Sol and Anthropic’s Claude Fable 5. Over the weekend, Alibaba previewed Qwen3.8, which it calls “one of the most powerful models available today” and “second only to Fable 5.” Both companies are releasing their models publicly — a stark contrast to the guarded approach of leading U.S. labs. Moonshot says Kimi K3 is the world’s largest open-source AI system, with 2.8 trillion parameters; Alibaba’s Qwen3.8 has 2.4 trillion. Neither OpenAI nor Anthropic disclose exact parameter counts.

Independent testing will determine whether these models live up to the claims. Moonshot plans to release full model weights on July 27; Alibaba says Qwen3.8 is “going open-weight soon.” The releases have already jolted the industry, echoing last year’s surprise from DeepSeek, which showed a low-cost model rivaling top U.S. systems. They also raise fresh questions about whether the billions U.S. companies spend on chips and data centers can secure a lasting advantage, especially as Washington tightens export controls and forces companies like Anthropic to pull capable systems from the market. The gap between U.S. and Chinese AI may be far narrower than it once seemed.

Source: The Verge