Anthropic: GLM-5.3 safeguards bypassed 64-100% of the time in cyber exploit tests
evilsocket · x · 2026-09-30
Anthropic's Frontier Red Team reports that Zhipu AI's GLM-5.3 matches Claude Mythos Preview in autonomously building end-to-end cyber exploits but ships without meaningful safeguards. In simulated tests, attackers bypassed its guardrails 64–100% of the time, while the same attacks failed against safeguarded Claude models. NIST's CAISI earlier called GLM-5.3 the most cyber-capable open-weight model to date. Anthropic had released its own cyber-capable model in a limited program that helped defenders find 10,000+ vulnerabilities, and warns GLM-5.3's lax safeguards materially expand attackers' capabilities even as they can benefit defenders.
More from Models
- Fable 5.2 Leads MazeBench Token Efficiency; Billion-Token Runs Underway — patience_cave · 2026-11-16
- Fable 5.2's Puzzle Skills Look Suspiciously Good, Reciting Outdated MazeBench Solution — patience_cave · 2026-11-16
- Fable 5.2 Tops MazeBench With 21% After 90 Hours and $6,000 in Credits — patience_cave · 2026-11-16
- Tencent Hy4 flash and InternLM Sage-S2.1-Flash both released, tool-calling tests requested — MajesticPuzzler · 2026-11-16
- User Shares Coral Reef Creation Labeled 'Qwen 4 Max (xhigh)', an Unannounced Model Name — intellectronica · 2026-11-16
- Stealth Model 'silicon-2' Surfaces in Devin, Suspected to Be Anthropic's Next Model — kimmonismus · 2026-11-16