Vulnerability Attack Analysis on WebGenBench Dashboards 1.0
100ASR (CodeQL)GPT-5.2
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| GPT-5.2Evaluation Protocol=Held-out (target-domain)2026.02 | 100 | 100 | 100 | 100 | |
| Claude-4.5 OpusEvaluation Protocol=Held-out (target-domain)2026.02 | 100 | 100 | 100 | 100 | |
| Gemini-3 ProEvaluation Protocol=Held-out (target-domain)2026.02 | 100 | 100 | 100 | 100 | |
| ComposerEvaluation Protocol=Held-out (target-domain)2026.02 | 100 | 100 | 100 | 100 | |
| GrokEvaluation Protocol=Held-out (target-domain)2026.02 | 100 | 75 | 100 | 75 | |
| GPT-5.2Evaluation Protocol=Cross-domain2026.02 | 78.57 | 83.33 | 76.47 | 69.57 | |
| Claude-4.5 OpusEvaluation Protocol=Cross-domain2026.02 | 75 | 81.25 | 75 | 80 | |
| ComposerEvaluation Protocol=Cross-domain2026.02 | 71.43 | 76.19 | 61.11 | 75 | |
| GrokEvaluation Protocol=Cross-domain2026.02 | 71.43 | 74.07 | 68.18 | 62.5 | |
| Gemini-3 ProEvaluation Protocol=Cross-domain2026.02 | 70 | 80 | 70 | 73.33 | |
| Gemini-3 FlashEvaluation Protocol=Cross-domain2026.02 | 55.56 | 72.73 | 55.56 | 61.54 | |
| Gemini-3 FlashEvaluation Protocol=Held-out (target-domain)2026.02 | 25 | 50 | 25 | 50 |