China’s cybersecurity giant 360 Security Technology has launched two AI‑powered tools it claims rival Anthropic’s Mythos vulnerability‑detection system, intensifying the US‑China race for offensive and defensive cyber capabilities. The move came just as Anthropic publicly accused Alibaba of orchestrating “the largest campaign ever to illicitly extract Claude’s capabilities,” an allegation that sent Alibaba shares to a 52‑week low.
At the ISC.AI 2026 conference in Beijing, founder Zhou Hongyi unveiled the pair under the name “Yitian Tulong” – “Heavenly Sword and Dragon Saber.” The first tool, Tulongfeng, automates software vulnerability discovery; Zhou called it “China’s version of Mythos.” The second, Yitianzhen, handles automated cyber defense and incident response. Zhou admitted that domestic AI models still lag US counterparts by 20–30% in base capability but argued that 360’s agent‑based approach – layering AI on decades of security expertise, vulnerability databases, and automation – delivers equivalent performance with less powerful models. “If Mythos is a top‑end chip, what we are building is a complete machine that can run stably, work 24 hours a day, and make fewer mistakes,” he said.
Zhou, who sits on China’s top political advisory body, framed the effort as a response to “one‑way transparency,” where US entities can probe Chinese infrastructure without reciprocity. 360 claims Tulongfeng has already found 3,432 vulnerabilities, 105 of which were confirmed by authorities; earlier AI‑driven methods caught about 1,000 flaws in systems such as Microsoft Office.
Meanwhile, Anthropic told the US Senate Banking Committee that between April 22 and June 5, 2026, Alibaba used 25,000 fake accounts in 28.8 million exchanges with Claude, targeting its advanced reasoning, coding, and multi‑step task skills. Anthropic described the operation as a “distillation campaign” that could speed China’s progress toward Mythos Preview‑level performance. The company had previously flagged similar efforts by DeepSeek, Moonshot, and MiniMax. Last week Washington ordered Anthropic to block foreign nationals from accessing Fable 5 and Mythos globally; the Pentagon also keeps Alibaba on a list of firms with alleged ties to the People’s Liberation Army.
The simultaneous revelations underscore a deepening AI arms race. Zhou’s announcement and the Alibaba allegations come amid years of mutual accusations over cyber espionage and tighter US chip export controls, which have narrowed China’s performance gap but not eliminated it.
