Pentesting Strategy Generation on Pentesting Scenarios (test)
73Strategy Success RateQwen-3-14B-GRPO
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen-3-14B-GRPOParameters=14B, Optimization=GRPO2026.05 | 73 | |
| Claude 4.5 Sonnet2026.05 | 65 | |
| GPT-52026.05 | 62 | |
| GPT 4.12026.05 | 60 | |
| Claude 3 Haiku2026.05 | 54 | |
| GPT-4o-mini2026.05 | 52 | |
| Gemini 2.5 Flash2026.05 | 45 | |
| Gemini 2.0 Flash2026.05 | 40 | |
| LLaMA-3.1-8BParameters=8B2026.05 | 40 | |
| Qwen-3-14BParameters=14B2026.05 | 39 | |
| GPT-3.5-turbo2026.05 | 36 |