The Wall Street Journal reports that China’s GLM-5.2 AI model can match Anthropic’s Mythos in cybersecurity tasks, signaling a rapidly narrowing AI gap.