China's Zhipu AI (Z.ai) has released its open-weight GLM-5.2, and some researchers are saying it can hold its own against Mythos in certain bug-finding and cybersecurity scenarios. Sure, GLM still lags behind Anthropic and OpenAI when it comes to more general tasks, but apparently China has managed to dramatically shrink the gap in AI capabilities between its models and those of the US. This is particularly troubling to the US government, which has been working hard to restrict China's access to powerful models like Mythos and Fable, as well as the hardware needed to train them. The Trump administration views Mythos and similar advanced AI models as serious national security threats - because nothing says "threat" like an AI that can find bugs in code. Meanwhile, OpenAI recently unveiled GPT-5.6, which has also raised concerns about potential misuse, leading to limited access. Because GLM is open-weight, anyone can download and run it on readily available hardware, giving power users deep access and flexibility. But it also makes it ripe for abuse by bad actors who can run it with little oversight - because what could go wrong with an open-source AI that's good at finding vulnerabilities?