Anthropic Says Zhipu Model GLM-5.3 Nears Claude Mythos Preview in Exploit Building
Anthropic's analysis shows Zhipu's open-weight GLM-5.3 model nearly matches Claude Mythos Preview in building cyber exploits while lacking effective safeguards.
Company
3 stories · Updated
Anthropic's analysis shows Zhipu's open-weight GLM-5.3 model nearly matches Claude Mythos Preview in building cyber exploits while lacking effective safeguards.
Chinese AI firm Z.ai has apologized after its ZCode tool uploaded user workspaces to cloud storage without permission and promised security improvements.
Chinese company Z.ai has released GLM 5.3, a powerful open-weight AI model capable of automating advanced coding and cybersecurity tasks.