Web security task completion on XBOW 104 tasks
91Task Completion RatePENTESTGPT V2
Evaluation results
| Method | Links | ||
|---|---|---|---|
| PENTESTGPT V2Model=Opus 4.5, Thinking Mode=Thinking2026.02 | 91 | — | |
| PENTESTGPT V2Model=GPT-5.2, Thinking Mode=Thinking2026.02 | 85 | — | |
| PENTESTGPT V2Model=Opus 4.5, Thinking Mode=Standard2026.02 | 81 | — | |
| PENTESTGPT V2Model=Gemini 3, Thinking Mode=Thinking2026.02 | 79 | — | |
| MAPTAModel=GPT-52026.03 | 76.9 | 190.8 | |
| PENTESTGPT V2Model=GPT-5.2, Thinking Mode=Standard2026.02 | 76 | — | |
| PENTESTGPT V2Model=Gemini 3, Thinking Mode=Standard2026.02 | 76 | — | |
| PentestAgentModel=GPT-5.2, Thinking Mode=Thinking2026.02 | 61 | — | |
| PentestAgentModel=Opus 4.5, Thinking Mode=Thinking2026.02 | 60 | — | |
| VulnBotModel=Opus 4.5, Thinking Mode=Thinking2026.02 | 58 | — | |
| VulnBotModel=GPT-5.2, Thinking Mode=Thinking2026.02 | 56 | — | |
| PentestGPTModel=Opus 4.5, Thinking Mode=Thinking2026.02 | 54 | — | |
| PentestAgentModel=Opus 4.5, Thinking Mode=Standard2026.02 | 54 | — | |
| PentestAgentModel=Gemini 3, Thinking Mode=Thinking2026.02 | 54 | — | |
| PentestGPTModel=GPT-5.2, Thinking Mode=Thinking2026.02 | 53 | — | |
| PentestAgentModel=GPT-5.2, Thinking Mode=Standard2026.02 | 52 | — | |
| AWEModel=Claude Sonnet 42026.03 | 51.9 | 53.1 | |
| AutoPTModel=Opus 4.5, Thinking Mode=Thinking2026.02 | 51 | — | |
| VulnBotModel=Gemini 3, Thinking Mode=Thinking2026.02 | 51 | — | |
| AutoPTModel=GPT-5.2, Thinking Mode=Thinking2026.02 | 50 | — | |
| VulnBotModel=Opus 4.5, Thinking Mode=Standard2026.02 | 50 | — | |
| PentestGPTModel=Gemini 3, Thinking Mode=Thinking2026.02 | 48 | — | |
| VulnBotModel=GPT-5.2, Thinking Mode=Standard2026.02 | 48 | — | |
| PentestGPTModel=Opus 4.5, Thinking Mode=Standard2026.02 | 47 | — | |
| PentestAgentModel=Gemini 3, Thinking Mode=Standard2026.02 | 46 | — | |
| PentestGPTModel=GPT-5.2, Thinking Mode=Standard2026.02 | 45 | — | |
| AutoPTModel=Gemini 3, Thinking Mode=Thinking2026.02 | 45 | — | |
| AutoPTModel=Opus 4.5, Thinking Mode=Standard2026.02 | 44 | — | |
| AutoPTModel=GPT-5.2, Thinking Mode=Standard2026.02 | 43 | — | |
| VulnBotModel=Gemini 3, Thinking Mode=Standard2026.02 | 43 | — | |
| PentestGPTModel=Gemini 3, Thinking Mode=Standard2026.02 | 41 | — | |
| AutoPTModel=Gemini 3, Thinking Mode=Standard2026.02 | 38 | — |