Vision-Language Multi-task Performance on MergeBench
32.6MMSI-Bench
Task-specific FT
Updated 1mo ago
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
2026.02 | 32.6 | 70.58 | 55.17 | 66.81 | 85 | 72.5 | 63.78 | |
2026.02 | 32.5 | 71.59 | 58.62 | 68.5 | 83.1 | 69.9 | 64.04 | |
2026.02 | 32.4 | 71.57 | 57.93 | 67.61 | 82.6 | 69.1 | 63.54 | |
2026.02 | 32.4 | 71.76 | 57.93 | 67.67 | 84.3 | 69.7 | 63.96 | |
2026.02 | 32.3 | 72.53 | 55.86 | 68.29 | 83.7 | 69.1 | 63.63 | |
2026.02 | 31.4 | 71.95 | 54.86 | 68.2 | 83.2 | 69 | 63.1 | |
2026.02 | 31.1 | 72.31 | 54.8 | 68.74 | 83.3 | 69.9 | 63.36 | |
2026.02 | 29.9 | 71.29 | 57.24 | 68.26 | 83.5 | 68.7 | 63.15 | |
2026.02 | 29 | 71.26 | 54.48 | 67.94 | 84.6 | 69.1 | 62.73 | |
2026.02 | 28.7 | 71.25 | 52.41 | 67.76 | 83.7 | 68.4 | 62.04 | |
2026.02 | 27.8 | 69.97 | 53.1 | 66.3 | 84.7 | 67.2 | 61.51 |