Open-weight GLM-5.2 nears frontier AI but lacks safety mitigations
SaferAI's report on Z.ai's GLM-5.2 found it refused none of the offensive cyber and bio tasks tested, while Claude Opus 4.7 refused so consistently that CyberGym couldn't be run. SaferAI's Henry Papadatos warns open weights can't be policed once downloaded.
Featured · Henry Papadatos
1 source
Daily brief
Get tomorrow's AI brief in your inbox
More stories today
- Scoble envisions AI agents mapping 3D printing businesses
- Tool enables AI agents to build fullstack apps from prompts
- Tencent elevates WorkBuddy as a top strategic AI priority
- Apple denies Qwen integration launched in China after guide pulled
- Depth Any Panoramas creates depth maps for panoramic imagery