CAPTCHA Solving on Xiaodun (test)
92.59AccuracyVIPER (R3)
Evaluation Results
| Method | Links | |
|---|---|---|
| VIPER (R3)Backend=gpt4o2026.01 | 92.59 | |
| VIPER (R3)Backend=grok-22026.01 | 92 | |
| Human2026.01 | 91.8 | |
| VRC-GraphNet2026.01 | 89.65 | |
| VIPER (R3)Backend=kimi2026.01 | 87.86 | |
| VIPER (R3)Backend=deepseek2026.01 | 87.76 | |
| Holistic2026.01 | 79.28 | |
| OedipusBackend=kimi2026.01 | 70.38 | |
| OedipusBackend=gpt4o2026.01 | 64.08 | |
| OedipusBackend=grok-22026.01 | 63.27 | |
| Ablation-R2Backend=gpt4o2026.01 | 62.31 | |
| Ablation-R2Backend=deepseek2026.01 | 60.24 | |
| Ablation-R2Backend=grok-22026.01 | 59.67 | |
| OedipusBackend=deepseek2026.01 | 59.53 | |
| Ablation-R2Backend=kimi2026.01 | 58.33 | |
| Ablation-R1Backend=kimi2026.01 | 12.96 | |
| Ablation-R1Backend=deepseek2026.01 | 11.82 | |
| Ablation-R1Backend=gpt4o2026.01 | 10.99 | |
| Ablation-R1Backend=grok-22026.01 | 10.46 |