Free Chinese AI models nearly match OpenAI's top secret cyber model in hacking test
Wired AI · rss · 2026-11-07
Wired reports on an experiment where AI models were tasked with breaking into simulated networks. The result: free Chinese models DeepSeek and GLM scored not far behind OpenAI's most advanced cyber offense model — which isn't even public.
Key takeaways:
- The test measured penetration/intrusion skills in simulated network environments
- Free, openly available Chinese models came close to OpenAI's top specialized model
- The finding highlights how fast frontier cyber capabilities are spreading via open models
More from Models
- 4B Nemotron 3 Nano claimed to beat frontier models on task-specific workflows — eliano · 2026-11-07
- Thinking Machines' Inkling 2 Small hits 81.7% on ARC-AGI-2, full weights coming next week — ziqiao_ma · 2026-11-07
- LithosAI claims fastest Kimi K3.1 Pro inference at 4.7 min per Intelligence Index task — JiaZhihao · 2026-11-07
- Applied Compute shares recipe for production RL with continual learning — ypatil125 · 2026-11-07
- Qwen3.8 27B writes 1,030-word replies, 6x longer than GPT-6 Sol's 174 — ArtificialAnlys · 2026-11-07
- China Telecom's Xing4.0-29B-A4B open-sourced: trained on Ascend NPUs, 75 on SWE-bench Verified — jacek2023 · 2026-11-07