Vulnerability Detection on Vulnerability evaluation dataset 1.0 (test)
97.2OpusD2 Dual-Pass
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| D2 Dual-PassDefense=Dual-Pass Analysis2026.02 | 97.2 | 96.1 | 94.5 | 97.2 | 95.5 | 35 | |
| D5 SAST Cross-RefDefense=SAST Cross-Referencing2026.02 | 97.2 | 96.9 | 96.6 | 97.2 | 96.6 | 47 | |
| D1 Comment StripDefense=Comment Stripping2026.02 | 96.9 | 93.2 | 89 | 95.9 | 91 | 30 | |
| D6 Comment AnomalyDefense=Comment Anomaly Detection2026.02 | 96.9 | 95.1 | 92.8 | 96.9 | 93.8 | 0 | |
| Baseline (C4+SP0)Defense=Baseline2026.02 | 94 | — | 93 | 91.9 | 92 | — |