Video Generation Reasoning on VideoThinkBench (test)
44.7Eyeballing-Point Accuracy
Sora-2
Updated 1mo ago
Evaluation Results
| Method | Links | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2025.11 | 44.7 | 38 | 34.5 | 81.9 | 51.9 | 57.5 | 1.3 | 13.3 | 40.4 | 68.6 | 65.3 | 61.2 | 79.1 | 68.6 | 49.8 | |
2025.11 | 36.2 | 26.3 | 50.5 | 80.1 | 69.9 | 82 | 5.3 | 0 | 43.8 | 90 | 86.3 | 65.6 | 82.3 | 81.1 | 56.2 | |
2025.11 | 33.6 | 24 | 32.5 | 98.5 | 66.7 | 85 | 0.5 | 0 | 42.6 | 97.2 | 85.2 | 69.6 | 80.6 | 83.2 | 56.1 | |
2025.11 | 27.8 | 21 | 34.5 | 94.9 | 83.7 | 67 | 1.9 | 0 | 41.3 | 94.8 | 84.5 | 66.7 | 83 | 82.3 | 55 |