GLM-5.3 completes 3 of 13 multi-stage cyberattack scenarios in Irregular's controlled tests
jcran · x · 2026-10-30
Controlled tests by security firm Irregular found that publicly downloadable GLM-5.3 completed 3 of 13 multi-stage attack scenarios and largely refused to decline offensive tasks, meaning chained cyberattacks are becoming feasible with models anyone can download.
Reposter seanhn highlights an asymmetry: attackers get capable models, while defenders—even those verified through Daybreak/CVP—hit refusals on patch analysis, bug finding, triage, and PoC development, pushing them toward ungated models with their own risks.
The takeaway: we can't count on model safeguards to prevent misuse, and defenders need to prepare for these capabilities spreading widely.
Related event: Security Tests Show GLM-5.3 Can Execute Multi-Stage Cyberattacks(3 posts)→
More from Models
- DeepSeek V4.1 Flash tops open-model usage among research agents, index data shows — AkariAsai · 2026-10-30
- Dev claims Gemini Argon makes 'researcher-level jumps' other models miss as 'carbon' looms — vitaliychiley · 2026-10-30
- Meta's Alexandr Wang touts Muse: 6.7x more concise than ChatGPT at answering the same question — alexandr_wang · 2026-10-30
- OpenAI's subsidy shift rewrites the score-vs-cost Pareto frontier, analysis finds — RexDouglass · 2026-10-30
- Phonic tops VAmoS benchmark for agentic speech-to-speech models — graceisford · 2026-10-30
- DeepSeek Frontend Leaks cyber-v3/v4 Keys; Still Hiring Attackers as V4 Excels at Security — teortaxesTex · 2026-10-30