Chinese cybersecurity firm 360 Security Technology announced the development of domestic artificial intelligence tools aimed at matching the capabilities of Anthropic’s Mythos, a US-developed system designed to detect software vulnerabilities. The announcement was made on Wednesday at the 2026 cybersecurity conference in Beijing by 360’s founder, Zhou Hongyi.

Mythos, which was previewed in April, has drawn significant attention for its ability to identify thousands of major vulnerabilities across operating systems, web browsers, and other software. However, cybersecurity experts have cautioned that such capabilities could potentially enhance offensive cyber operations. In response to national security concerns, the US government recently ordered Anthropic to suspend exports of a less powerful version of Mythos worldwide and to all foreign nationals.

Zhou introduced two AI tools under the umbrella “Yitian Tulong,” meaning “Heavenly Sword and Dragon Saber,” named after a classic Chinese martial arts novel. The first tool, “Tulongfeng,” is designed for automatic software vulnerability discovery and was described by Zhou as “China’s version of Mythos.” The second system, “Yitianzhen,” focuses on automating cyber defense and incident response. Zhou characterized vulnerability-finding AI as a strategic national asset capable of both defending critical infrastructure and providing offensive advantages.

The development comes amid ongoing tensions between China and the US concerning cyber operations targeting critical infrastructure, with both nations accusing each other of offensive cyber activities. Zhou emphasized the risk of “one-way transparency,” warning that if US entities can utilize Mythos-like AI to scan software and systems while Chinese companies are denied similar tools, China would be at a strategic disadvantage.

According to 360’s claims, Tulongfeng has identified 3,432 software vulnerabilities, 105 of which have been confirmed by Chinese authorities, though these figures have not been independently verified. Zhou acknowledged that Chinese AI models still lag behind US counterparts, citing a 20 to 30 percent gap in base capability largely due to US export controls on high-end computing chips since 2022. These restrictions are intended to limit China’s military AI advancements.

Rather than relying solely on high-end computing power, Zhou explained that 360 is pursuing an “agent” approach, integrating AI models with security expertise, vulnerability databases, and automated tools. This strategy aims to create a comprehensive and stable system capable of continuous operation with a low error rate. Zhou contrasted this with the US approach, saying, “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 also described 360’s method as organizing a professional team for cyber offense and defense, as opposed to cultivating individual “genius hackers.”

Experts have noted that vulnerabilities in AI systems have already been exploited globally. Anthropic reported that hackers used weaknesses in its Claude AI to attack approximately 30 organizations last year, while a separate industry study found that 67 percent of executives surveyed had faced AI-driven cyberattacks within the past year.

Zhou, a respected figure in China’s technology sector and a member of the country’s top political advisory body, founded 360 Security Technology, which is well-known domestically for its antivirus software and broader cybersecurity services for enterprises and government entities. The company’s new AI tools mark a significant step in China’s efforts to develop domestic cyber capabilities amid increasing competition and restrictions from the US.