Inside Artificial Analysis Cyber Index: IBM, NVIDIA, Vercel Build Cyber Defense Eval Standard
ArtificialAnlys · x · 2026-11-13
Artificial Analysis details its Cyber Index methodology: it measures cyber defense capability — discovering vulnerabilities, reproducing/validating them, and patching without breaking functionality — via three evaluations (CWE-Bench-AA from Collinear AI's CWE-bench, DeepsecBench-AA from Vercel, CyberGym-E2E-AA from Berkeley RDI), all independently run.
The index is built with the Cyber Index Alliance (Collinear AI, IBM, NVIDIA, Pasta Lab, Vercel). Leaderboard: Kimi K3.1 Pro tops at 64, Tencent Hy4 at 58, DeepSeek V4.1 Flash 1017 (Max) at 57. Lowest cost per task: GPT-6 Luna (Max) at $0.12, MiMo-V2.6-Pro at $0.18, Beam 1.0 at $0.20.
Related event: Tencent Hunyuan Hy4 Ranks Third in New Cyber Defense Index(2 posts)→
More from Models
- Daily AI brief: Gemini 4 Hydrogen pricing revealed, Anthropic to list as ANT on Nasdaq Nov 19 — testingcatalog · 2026-11-13
- After two months daily-driving Claude Code, one reviewer says Opus 5.5 beats GPT everywhere except computer use — kimmonismus · 2026-11-13
- HF open-sources Ettin rerankers: six sizes, SOTA, full training recipe — antoine_chaffin · 2026-11-13
- LightOn open-sources DenseOn: SOTA single and multi-vector retrieval models — antoine_chaffin · 2026-11-13
- Voice Code Bench adds 8 ASR models as ElevenLabs Scribe V3 takes the top spot — lukeharries · 2026-11-13
- Claude Hit by Elevated Errors Across claude.ai, Claude Code and API Platform — ClaudeAI-mod-bot · 2026-11-13