China’s 360 Security unveils AI tools that “match” Anthropic’s Mythos

360 Security Technology, one of China’s largest cybersecurity firms, has announced two AI-powered security tools with claims that they can rival Anthropic’s Mythos vulnerability-detection system, stating that the American model is a strategic threat Beijing must answer to.

Founder Zhou Hongyi introduced the tools at the ISC.AI 2026 cybersecurity conference in Beijing under the collective name “Yitian Tulong,” a reference to a classic Chinese martial arts novel translated as “Heavenly Sword and Dragon Saber,” according to Reuters.

The firstof the two tools, Tulongfeng, automates software vulnerability discovery. Zhou went ahead to call it “China’s version of Mythos.” The second AI tool, Yitianzhen, handles automated cyber defense and incident response. Together, 360 is positioning them as both shield and sword in an escalating AI-driven security arms race.

Mythos’ foundations

In Anthropic’s April Mythos preview, the AI company stated that it was a system capable of autonomously finding flaws in widely used software. The company also reported at the time that an early version had uncovered “thousands” of serious vulnerabilities across multiple operating systems, web browsers, and other applications.

The system’s dual-use potential triggered an immediate alarm, with cybersecurity experts warning that the same defensive capabilities could significantly arm offensive operations. Washington responded by ordering Anthropic to suspend exports of a less powerful Mythos variant to all foreign destinations and nationals, citing national security.

360 Security pitches agents over compute

Zhou acknowledged that Chinese AI models still trail their American counterparts, saying, “Objectively speaking, domestic models still have a 20%-30% gap in base capability.”

His argument, however, was that China cannot afford to wait for that gap to close. Instead, 360 took an “agent” approach as described by Zhou, with a focus on layering AI models on already existing security expertise, vulnerability databases, and automated tooling. He claimed this combination gives Tulongfeng capabilities equivalent to Mythos despite using less powerful underlying 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,” Zhou said, according to Reuters. “If the U.S. route is to cultivate a genius hacker, 360’s route is to organise a professional attack-and-defence team.”

360 claimed Tulongfeng had identified 3,432 software vulnerabilities so far, with 105 confirmed by Chinese authorities. Earlier in April, the company reported that its AI-driven methods detected about 1,000 vulnerabilities in systems including Microsoft Office.

China’s potential “one-way transparency” problem

Without domestic equivalents to Mythos, Zhou argued that China faces a “one-way transparency” problem where American entities possess the capacity to probe Chinese software and critical infrastructure, with Chinese firms unable to do the same.

“This kind of powerful weapon that can change the landscape of cyber offence and defence cannot be held only by others,” Zhou said, according to 360’s published transcript.

Zhou sits on China’s top political advisory body, which makes his remarks carry additional political weight. His warnings echo coverage from Chinese state media describing Mythos as demonstrating “unprecedented cyberattack capabilities.”

The announcement arrives amid years of mutual accusations between Washington and Beijing over offensive cyber operations targeting critical infrastructure. The enforcement of tighter U.S. advanced chip exports since 2022 has led to constraints in Chinese AI development; however, this performance gap has recently narrowed.