China Has Its Own Mythos Now, Says Qihoo 360 Founder. And One Version Is Free
The U.S. has restricted Anthropic’s Claude Mythos through export controls and a vetted partner program, excluding Chinese firms. In response, Qihoo 360 CEO Zhou Hongyi said China needs its own equivalent and unveiled Tulong Feng, an AI vulnerability agent, plus Yitian Zhen and a domestic security coalition called Panshi Zhidun. Zhou described Mythos as “cyber nuclear weapons” and claimed Tulong Feng has found 3,432 vulnerabilities, including 105 confirmed by regulators. Separately, Beijing-based Z.ai released GLM-5.2 under the MIT license, with no access restrictions. Its cybersecurity performance was highlighted: 39% on Semgrep’s IDOR detection test, ahead of Claude Code, and near parity with Claude Opus 4.8 on a capture-the-flag benchmark, at much lower cost per finding. Z.ai framed the release as a push for universal access to advanced security models.
