Multi-view spatial reasoning on MMSI-Bench
42.7Overall Accuracy
Gemini-3-Pro
Updated 24d ago
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
2026.07 | 42.7 | 42.9 | 50 | 38 | 40.9 | |
2026.07 | 42 | 42.9 | 47.6 | 36 | 40.4 | |
2026.07 | 35.5 | 41.2 | 35.4 | 28.7 | 25.8 | |
2026.07 | 31.7 | 29.3 | 33.8 | 38.7 | 31.3 | |
2026.07 | 31.7 | 36.6 | 22.3 | 27.3 | 28.4 | |
2026.07 | 31.6 | 36.2 | 31.5 | 24.7 | 24.7 | |
2026.07 | 30.1 | 31 | 26.9 | 31.3 | 28.8 | |
2026.07 | 29.8 | 35.9 | 26 | 21 | 23.2 | |
2026.07 | 29.6 | 32.6 | 31.5 | 22.7 | 25.8 | |
2026.07 | 27.7 | 34.2 | 32.3 | 26 | 23.7 | |
2026.07 | 27.1 | 28.4 | 21.5 | 29.3 | 25.8 | |
2026.07 | 27.1 | - | - | - | - | |
2026.07 | 26.9 | 29.1 | 29.4 | 21.9 | 22.5 | |
2026.07 | 26.1 | 28.5 | 25.4 | 18 | 26.3 | |
2026.07 | 25.7 | 30.8 | 23.8 | 24.6 | 14.6 | |
2026.07 | 25.4 | 30.3 | 23.3 | 16 | 21.2 | |
2026.07 | 23.9 | 23.9 | 24.6 | 18.7 | 25.3 |