AnalysisPolicySeptember 29, 2026

Anthropic: GLM-5.3 nears Claude Mythos on cyber exploits

Read original source →anthropic.com

Anthropic's Frontier Red Team says Zhipu's open-weight GLM-5.3 builds full control flow hijacks in 4% of Binary Exploitation benchmark trials versus 6% for Claude Mythos Preview, while earlier models like Claude Opus 4.6 and GLM-5.2 scored zero. Attackers bypass GLM-5.3's safeguards 64%-100% of the time in simulated tests.

2 sources

More stories today

Open the live feed