| object-hallucination-evaluation-on-pope | NEO-9B | #6 | F1: 88.4 |
| infographic-visual-question-answering-on-infovqa | NEO-2.2B | #7 | ANLS: 63.2 |
| visual-question-answering-vqa-on-hallusionbench-average-accuracy | NEO-9B | #20 | Accuracy: 46.4 |
| visual-question-answering-on-docvqa-test | NEO-2.2B | #45 | ANLS: 0.899 |
| optical-character-recognition-ocr-on-ocrbench | NEO-9B | #57 | Score: 777 |
| visual-question-answering-on-mm-vet | NEO (Qwen3-8B) | #57 | GPT-4 score: 53.6 |
| visual-question-answering-on-mm-vet | NEO (Qwen3-1.7B) | #69 | GPT-4 score: 49.6 |
| chart-question-answering-on-chartqa | NEO-9B | #59 | 1:1 Accuracy: 82.1 |
| vision-language-understanding-on-mmstar | NEO-9B | #76 | Accuracy: 62.4 |
| multimodal-reasoning-on-mmmu | NEO-9B | #77 | Accuracy: 54.6 |