When the US government restricted access to Anthropic’s cybersecurity AI model Mythos behind export controls, China’s response was swift. At ISC.AI 2026 in Beijing last week, Qihoo 360 founder Zhou Hongyi took the stage and unveiled Tulong Feng, calling it China’s answer to Mythos.
Zhou did not mince words. He called Mythos cyber nuclear weapons in the AI age. His pitch was direct: US organizations can use Mythos to scan your vulnerabilities, but you do not even have the right to look at it. Chinese companies are excluded from Anthropic’s Glasswing partner program. So Zhou built an alternative. Alongside Tulong Feng, he announced Yitian Zhen, an automated defense platform, and a domestic security coalition called Panshi Zhidun.
Tulong Feng has reportedly found 3,432 vulnerabilities, with 105 confirmed by Chinese regulators and several flagged as high-severity. Zhou claimed an agent-first approach, coordinating specialized models rather than relying on a single frontier system, offsets whatever base-model gap still exists with US rivals.
Then Z.ai, also known as Zhipu AI, went further. The Beijing lab released GLM-5.2 under an MIT license. No subscription gates. No geographic restrictions. Free and modifiable by anyone. The timing was pointed: it dropped shortly after the US pulled Mythos 5 and Fable 5 offline for foreign nationals.
The cybersecurity benchmarks are notable. On Semgrep’s insecure direct object reference detection test, GLM-5.2 scored 39% on the F1 metric, ahead of Claude Code. A separate Graphistry capture-the-flag evaluation showed it matching Claude Opus 4.8. Cost per finding: roughly $0.17, versus over $1 for Claude-based workflows.
Z.ai co-founder Tang Jie called Anthropic’s withdrawal deeply regrettable. Technical lead Qinkai Zheng was more direct: We want the model accessible to everyone. When Elon Musk predicted China would not match Fable-level capability until Q1 2027, Tang replied: Won’t take that long.
