Chinese AI Matches Anthropic's Mythos in Bug Detection

A Chinese AI startup's model, GLM-5.2, can now match Anthropic's Mythos in finding software security vulnerabilities. This comes from a new report in The Wall Street Journal. What's interesting is that GLM-5.2 performs on par with Mythos in identifying software bugs. This capability is crucial for patching vulnerabilities before hackers exploit them. The model is also open-source, allowing anyone to download and modify it. This flexibility is attractive to businesses but also raises concerns that cybercriminals could use it for harmful purposes. The report notes that China has not overtaken the U.S. in overall AI. However, in cybersecurity, this performance gap has narrowed significantly. The Wall Street Journal cites benchmark data showing GLM-5.2 has even outperformed Claude Opus 4.8 in some security evaluations. Researchers say additional prompting helps it reach Mythos-level bug-finding performance. The bottom line is that the gap in AI capabilities is closing rapidly, especially in critical areas like cybersecurity.