Tag / 1 article

z-ai

1. Articleglm-5-3Oct 3, 2026

   Anthropic: open-weight GLM-5.3 builds exploits, lacks safeguards

   Anthropic says Z.ai's open-weight GLM-5.3 nearly matches Claude Mythos Preview at building cyber exploits, and its safeguards are easy to bypass.

   3 min read