Vision-Language Understanding on MMBench (test)
78.8Overall AccuracyBaseline
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| BaselineModel=Qwen2.5-VL-3B, Authorization Status=unauthorized2025.10 | 78.8 | 81.3 | 83.1 | 69.2 | 84.9 | 66.5 | 75.4 | |
| BaselineModel=InternVL3-1B, Authorization Status=authorized2025.10 | 72.7 | 77.4 | 82.4 | 58.7 | 79.4 | 50.3 | 66.8 | |
| NTEModel=InternVL3-1B, Authorization Status=authorized2025.10 | 72.6 | 77.8 | 82 | 58.3 | 78.9 | 50.9 | 67.3 | |
| NTEModel=Qwen2.5-VL-3B, Authorization Status=unauthorized2025.10 | 18.3 | 29.2 | 15.6 | 17 | 12.8 | 17.3 | 22.3 |