Multiple Choice Visual Question Answering on NegBench VOC 2007 (test)
95.37AccuracyInternVL3.0-8B
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| InternVL3.0-8BParameters=8B2026.03 | 95.37 | 1 | 84.24 | 1 | 96.84 | |
| InternVL3.5-8BParameters=8B2026.03 | 92.21 | 2 | 80.88 | 3 | 97.5 | |
| InternVL2.5-8BParameters=8B2026.03 | 92.01 | 3 | 80.76 | 4 | 98.09 | |
| Qwen2.5-VL-7BParameters=7B2026.03 | 91.97 | 4 | 83.74 | 2 | 92.23 | |
| Gemma-3-12BParameters=12B2026.03 | 87.5 | 5 | 80.06 | 5 | 87.52 | |
| Qwen2.0-VL-7BParameters=7B2026.03 | 85.35 | 6 | 78.22 | 6 | 87.64 | |
| Qwen3.0-VL-8BParameters=8B2026.03 | 84.08 | 7 | 76.88 | 7 | 84.08 | |
| LLaVA-1.5-13BParameters=13B2026.03 | 77.38 | 8 | 71.12 | 8 | 77.38 | |
| LLaVA-1.6-13BParameters=13B2026.03 | 65.28 | 9 | 60.96 | 9 | 68.97 | |
| InternVL2.0-8BParameters=8B2026.03 | 58.68 | 10 | 51.98 | 10 | 93.52 | |
| QwenVL-7B-ChatParameters=7B2026.03 | 23.71 | 11 | 23.14 | 11 | 55.6 |