智谱GLM-5.2漏洞发现能力可比肩Mythos

Odaily Planet Daily reports that domestic AI model company Zhipu AI recently released its open-weight model GLM-5.2. Some researchers claim that GLM-5.2 is comparable to Anthropic's Mythos model in certain vulnerability detection and cybersecurity scenarios. Data from cybersecurity firm Semgrep shows that in some benchmark tests, GLM-5.2 outperformed Anthropic's Claude Opus 4.8 model released in May. Researchers pointed out that after receiving further instructions, both Opus 4.8 and GLM-5.2 are comparable to Mythos in vulnerability discovery capabilities. As an open-weight model, GLM-5.2 can be downloaded by anyone and run on off-the-shelf hardware.