The AI arms race just got a cybersecurity subplot. 360 Security Technology, one of China’s largest cybersecurity firms, has launched two new AI systems designed to compete directly with Anthropic’s Claude Mythos model, which has been turning heads since its preview release in April 2026.
The systems, named Tulongfeng and Yitianzhen, were unveiled on June 24. Tulongfeng focuses on automated software vulnerability discovery, while Yitianzhen handles cyber defense and incident response.
What Tulongfeng actually does
Tulongfeng uses AI to identify weaknesses in software at machine speed. The numbers so far are notable. Tulongfeng has already identified over 3,400 vulnerabilities in software systems, according to 360 Security. That puts it in roughly the same ballpark as Anthropic’s Mythos, which identified thousands of vulnerabilities across various software applications following its April preview release.
360 Security previously showcased its Multi-Agent Collaborative Vulnerability Discovery System during the Tianfu Cup hacking contest, China’s answer to the well-known Pwn2Own competition. That earlier system laid the groundwork for what Tulongfeng has become.
