Discriminative QA on MMAD
89.37Industrial Components ScoreMAU-GPT
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| MAU-GPTType=Specialized Model, Number of Parameters=4B2026.01 | 89.37 | 53.73 | 73.75 | 77.56 | 74.75 | 86.81 | 82.12 | |
| UIO2-LargeType=Generalist Model, Number of Parameters=1B2026.01 | 66.61 | 53.26 | 53.28 | 57.06 | 61.61 | 66.4 | 59.7 | |
| Gemma-3Type=Generalist Model, Number of Parameters=4B2026.01 | 63.06 | 58.12 | 43.75 | 48.64 | 51.5 | 33.86 | 42.45 | |
| Yi-VLType=Generalist Model, Number of Parameters=6B2026.01 | 62.39 | 60.05 | 58.54 | 47.96 | 52.91 | 83.17 | 73.54 | |
| AnomalyGPTType=Specialized Model, Few-shot settings=4-shot2026.01 | 58.14 | 69.78 | 81 | 66.44 | 77.41 | 43.39 | 66.03 | |
| AnomalyGPTType=Specialized Model, Few-shot settings=3-shot2026.01 | 57.04 | 69.74 | 80.19 | 67.14 | 76.62 | 42.92 | 65.61 | |
| AnomalyGPTType=Specialized Model, Few-shot settings=2-shot2026.01 | 56.2 | 70.16 | 79.46 | 64.38 | 75.79 | 41.25 | 64.54 | |
| UIO2-XlType=Generalist Model, Number of Parameters=3B2026.01 | 55.55 | 55.17 | 64.04 | 53.28 | 60.38 | 67.84 | 59.38 | |
| AnomalyGPTType=Specialized Model, Few-shot settings=1-shot2026.01 | 50.93 | 66.07 | 78.7 | 60.41 | 79.31 | 39.76 | 62.53 | |
| InternVL-2.5Type=Generalist Model, Number of Parameters=8B2026.01 | 46.58 | 57.77 | 47.71 | 49.66 | 55.31 | 41.62 | 45.31 | |
| LLaVA-1.5Type=Generalist Model, Number of Parameters=7B2026.01 | 40.1 | 30.21 | 39.46 | 50.66 | 45.83 | 55.4 | 43.61 | |
| LLaVA-1.5Type=Generalist Model, Number of Parameters=13B2026.01 | 11.57 | 27.31 | 25.26 | 32.13 | 27.76 | 15.24 | 23.21 |