Home / Technology / China's AI rivals US "Mythos" with Yitian Tulong
China's AI rivals US "Mythos" with Yitian Tulong
25 Jun
Summary
- China's 360 Security launched two AI models for cyber defense.
- One AI is called 'Chinese Mythos' and found many software flaws.
- The company acknowledges a capability gap but prioritizes team building.
At the ISC.AI 2026 conference on June 24, 2026, China's 360 Security Technology introduced "Yitian Tulong," a suite of two AI models designed for vulnerability discovery and automated defense. The model named Tulongfeng is positioned as China's response to advanced US AI systems capable of identifying and exploiting software flaws.
While Tulongfeng reportedly discovered 3,432 software vulnerabilities, with 105 confirmed by the Chinese government, its founder Zhou Hongyi noted a 20-30% capability gap compared to leading US models. Instead of solely focusing on catching up in raw capability, 360 Security is prioritizing the development of a comprehensive, professional attack-and-defense team.
This strategic approach aims to create a stable, 24-hour operational cyber defense capability, drawing a parallel between the US focus on a "genius hacker" and China's emphasis on organized, collaborative teamwork for robust cybersecurity.
