CAPTCHA Solving on DingX (test)
95.4AccuracyHuman
Evaluation Results
| Method | Links | |
|---|---|---|
| Human2026.01 | 95.4 | |
| VIPER (R3)Backend=gpt4o2026.01 | 91.48 | |
| VIPER (R3)Backend=kimi2026.01 | 89.2 | |
| Holistic2026.01 | 88.52 | |
| VIPER (R3)Backend=deepseek2026.01 | 87.88 | |
| VIPER (R3)Backend=grok-22026.01 | 86 | |
| VRC-GraphNet2026.01 | 85.34 | |
| Ablation-R2Backend=grok-22026.01 | 74.82 | |
| OedipusBackend=kimi2026.01 | 74.63 | |
| Ablation-R2Backend=gpt4o2026.01 | 73.98 | |
| OedipusBackend=gpt4o2026.01 | 73.16 | |
| OedipusBackend=grok-22026.01 | 70.94 | |
| Ablation-R2Backend=deepseek2026.01 | 70.46 | |
| Ablation-R2Backend=kimi2026.01 | 68.24 | |
| OedipusBackend=deepseek2026.01 | 68.14 | |
| Ablation-R1Backend=grok-22026.01 | 26.38 | |
| Ablation-R1Backend=kimi2026.01 | 24.97 | |
| Ablation-R1Backend=gpt4o2026.01 | 24.87 | |
| Ablation-R1Backend=deepseek2026.01 | 23.12 |