Multiple Choice Questions on AgMMU
77.9Accuracy (Disease)GPT-o4-mini
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| GPT-o4-miniSize=-2026.03 | 77.9 | 85.4 | 90.3 | 93.8 | 84.3 | 86.5 | |
| Gemini 1.5 ProSize=-2026.03 | 76.2 | 81.1 | 82.8 | 88.1 | 76.9 | 82.4 | |
| AgroNVILASize=8B, Evaluation protocol=30% Data2026.03 | 70 | 74.3 | 69.7 | 85.3 | 89.9 | 78.6 | |
| NVILA-Lite-8BSize=8B2026.03 | 68.6 | 70.8 | 71 | 84.1 | 79.9 | 75.4 | |
| AgroNVILASize=8B, Evaluation protocol=10% Data2026.03 | 67.9 | 77.1 | 67.6 | 85.9 | 84.9 | 77.3 | |
| LLaMA-3.2Size=11B2026.03 | 66.2 | 75 | 78.6 | 89.6 | 79.9 | 78.3 | |
| LLaVA-OneVisionSize=8B2026.03 | 65.7 | 72.9 | 71.2 | 85.9 | 78.8 | 75.4 | |
| CambrianSize=8B2026.03 | 65 | 70.1 | 59.3 | 79.1 | 86 | 72.8 | |
| LLaVA-1.5Size=13B2026.03 | 64.7 | 67.4 | 65.5 | 80.4 | 73.7 | 70.8 | |
| Claude 3 HaikuSize=-2026.03 | 62.1 | 71.2 | 52.8 | 81.5 | 52 | 63.8 | |
| AgroNVILASize=8B, Evaluation protocol=Zero-shot2026.03 | 57.9 | 66 | 62.8 | 77.3 | 58.7 | 64.6 |