Chinese company Z.ai claims it can match Mythos in cybersecurity

China Zhipu AI (Z.ai) released the open-weight GLM-5.2, and some researchers have claimed that it is Corresponds to Mythos In some error detection and cybersecurity scenarios. While GLM lags behind models from Anthropic and OpenAI In other, more general tasks, China appears to have largely succeeded in narrowing the capabilities gap between its models and those of the United States.

This level of progress is of particular concern to the US government, which has worked to restrict China’s access to powerful models such as the anthropic model. Mythology and Fable, as well as the hardware needed to train and operate them. The Trump administration views Mythos and other advanced AI models capable of identifying vulnerabilities as serious threats to national security. OpenAI was recently unveiled GPT-5.6which has also raised concerns about its potential for misuse and limited accessibility.

Since the GLM is an open-weighted model, it can be Downloaded and played by anyone On readily available devices. This gives it great flexibility and allows experienced users deep access, but also makes it vulnerable to abuse by bad actors who can operate it with little oversight.