A Chinese cybersecurity firm says it has developed homegrown artificial intelligence tools designed to rival Anthropic’s Mythos system — a U.S.-developed program that has raised alarms across the global cybersecurity community.
360 Security Technology made the announcement Wednesday, framing Anthropic’s Mythos as a strategic cyber weapon that China simply cannot allow to go unanswered.
Mythos, which was previewed in April, is an AI system capable of detecting vulnerabilities in software. Cybersecurity experts have warned the technology could significantly amplify the threat of cyberattacks. This month, the U.S. government ordered Anthropic to halt exports of a less powerful version of the program, citing concerns about national security.
At the ISC.AI 2026 cybersecurity conference held in Beijing, 360 founder Zhou Hongyi took the stage to introduce two new AI security tools grouped under the name “Yitian Tulong” — a reference to a beloved Chinese martial arts novel, with the phrase translating roughly to “Heavenly Sword and Dragon Saber.”
Zhou explained that one of the tools, called “Tulongfeng,” is built to automatically find software vulnerabilities, and he referred to it as “China’s version of Mythos.” The second tool, “Yitianzhen,” is designed to automate cyber defense and incident response.
“This kind of powerful weapon that can change the landscape of cyber offence and defence cannot be held only by others,” Zhou said in remarks published by 360.
He characterized vulnerability-detecting AI as a national strategic asset, one that could serve both defensive and offensive purposes when it comes to critical infrastructure.
China and the United States have long traded accusations over offensive cyber operations targeting each other’s critical systems. The unveiling by 360 represents the most prominent Chinese response yet to the concerns sparked by Mythos.
Anthropic stated in April that its Mythos Preview tool had uncovered “thousands” of significant vulnerabilities across operating systems, web browsers, and other software. The U.S. government subsequently directed the company to suspend exports of a scaled-down version of Mythos to all international destinations and foreign nationals, again citing national security.
Zhou warned that China risked falling into a situation of “one-way transparency” — where U.S. entities could use Mythos-like systems to scan Chinese software and infrastructure, while Chinese companies had no comparable tools to do the same. His comments reflect a broader anxiety in China, with state media describing Mythos as displaying “unprecedented cyberattack capabilities.” Zhou himself sits on China’s top political advisory body.
360 reported that Tulongfeng has identified 3,432 software vulnerabilities, including 105 that have been confirmed by Chinese authorities. Reuters was unable to independently verify those figures.
Zhou said his company would not simply replicate the American approach, which he described as depending on “the strongest model, the strongest computing power and the strongest chips.” U.S. export restrictions on advanced chips, tightened since 2022, have limited China’s ability to keep pace with American AI developers, though that gap has reportedly narrowed over the past year. The U.S. has defended those restrictions by arguing the chips could be used to enhance the Chinese military’s AI capabilities.
“Objectively speaking, domestic models still have a 20%-30% gap in base capability,” Zhou acknowledged. “China cannot wait until model capabilities have fully caught up before starting vulnerability discovery, because we cannot afford to wait.”
Instead, Zhou said 360 is pursuing what he called an “agent” approach — combining AI models with security expertise, vulnerability databases, and automated tools. He claimed this strategy, which he said only 360 has successfully deployed, gives Tulongfeng capabilities on par with Mythos.
“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. “If the U.S. route is to cultivate a genius hacker, 360’s route is to organise a professional attack-and-defence team.”
The announcement comes against a backdrop of growing AI-related cyber threats. Anthropic previously disclosed that hackers exploited weaknesses in its Claude AI to attack roughly 30 organizations worldwide. A separate study by IBM and Palo Alto Networks found that 67% of 1,000 executives surveyed said they had been targeted by AI-driven attacks within the past year.
Zhou is a well-known figure in China’s technology sector. He founded 360, which built its reputation through antivirus software before expanding into cybersecurity services for businesses and government clients.
