GDPx: Cybersecurity
Cybersecurity · 2026-11-01
GDPx's Cybersecurity domain: 18 tasks from Arithmetic on finding, exploiting and fixing proprietary zero-days inside live, multi-service enterprise environments. Scores are pass@1 over three attempts per task. A component of the GDPx score, which weights its 11 domains equally.
Top models (higher is better)
| Model | Score |
|---|---|
| Muse Wonder Full | 35.2 |
| Muse Wonder | 24.1 |
| Gemini 4 Argon | 18.5 |
| Gemini 4 Carbon | 16.7 |
| Grok 4.7 | 14.8 |
| GPT-6 Luna | 11.1 |
| Kimi K3 | 11.1 |
| Claude Haiku 5.5 | 9.3 |
| Claude Opus 5.5 | 9.3 |
| GLM-5.3 | 7.4 |
| MiniMax M3.1 Flash | 7.4 |
| Claude Fable 5.1 | 3.7 |
| Claude Sonnet 5.5 | 3.7 |
| GPT-6.1 Sol | 0.0 |
| GPT-6.1 Terra | 0.0 |
| GPT-6 Astra | 0.0 |
| Muse Spark Open | 0.0 |