China's Z.ai GLM-5.2 claims to match Mythos in cybersecurity, lags on general tasks
China's Zhipu AI has released its open-weight model GLM-5.2, which researchers say matches Mythos in certain bug-finding and cybersecurity scenarios, though it still lags behind Western models on general tasks.
China's Zhipu AI (Z.ai) has released its latest open-weight model, GLM-5.2. According to some researchers, the model can match Anthropic's Mythos in bug-finding and cybersecurity scenarios. While GLM still lags behind models from Anthropic and OpenAI on general tasks, China has significantly narrowed the capability gap with US models.
This advancement is particularly concerning to the US government, which has sought to restrict China's access to powerful models like Anthropic's Mythos and Fable, as well as the hardware needed to train and run them. The Trump administration views Mythos and other advanced AI models capable of identifying vulnerabilities as serious national security threats. Recently, OpenAI unveiled GPT-5.6, which has also raised concerns about potential misuse and has limited access.
Because GLM is an open-weight model, it can be downloaded and run by anyone on readily available hardware. This provides great flexibility and deep access for power users, but also makes it ripe for abuse by bad actors who can run it with little oversight.
