Video Anomaly Detection on CaC-Bench Main
98.2Recall (Anomalous)Skyra-RL
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| Skyra-RL2026.05 | 98.2 | 50 | 66.3 | 1.8 | 50 | 3.5 | 50 | |
| CaC2026.05 | 91 | 76.7 | 83.3 | 72.4 | 88.9 | 79.8 | 81.7 | |
| Gemini-2.5-Pro2026.05 | 85.4 | 53 | 65.4 | 24.4 | 62.6 | 35.1 | 54.9 | |
| UnifiedReward2026.05 | 53.8 | 56.3 | 55 | 58.2 | 55.7 | 56.9 | 56 | |
| Gemini-3.1-Pro2026.05 | 53.7 | 69.2 | 60.5 | 58.3 | 47.6 | 52.4 | 56 | |
| UnifiedReward-Think2026.05 | 42.8 | 50.4 | 46.3 | 57.8 | 50.3 | 53.8 | 50.3 | |
| VideoVeritas2026.05 | 41.4 | 53.9 | 46.8 | 64.6 | 52.4 | 57.9 | 53 | |
| VR-Thinker2026.05 | 38.2 | 55.4 | 45.2 | 69.2 | 52.8 | 59.9 | 53.7 | |
| GPT-4o2026.05 | 36.4 | 45.4 | 40.4 | 56.2 | 46.9 | 51.1 | 46.3 | |
| Qwen3-VL-8BParameters=8B2026.05 | 10.8 | 57.4 | 18.2 | 92 | 50.8 | 65.4 | 51.4 | |
| Qwen3-VL-32BParameters=32B2026.05 | 6.6 | 70.2 | 12.1 | 97.2 | 51 | 66.9 | 51.9 |