| chart-question-answering-on-chartqa | LFM (w/ Qwen2.5-VL) | – | 1:1 Accuracy: 89.2 |
| chart-question-answering-on-charxiv-dq | LFM (w/ Qwen2.5-VL) | – | Accuracy: 75.8 |
| chart-question-answering-on-charxiv-rq | LFM (w/ Qwen2.5-VL) | – | Accuracy: 44.7 |
| multimodal-reasoning-on-mathvista | LFM (w/ Qwen2.5-VL) | – | Accuracy: 69.8 |
| vision-language-understanding-on-mmbench-v1-1 | LFM (w/ Qwen2.5-VL) | – | Accuracy: 84.2 |
| vision-language-understanding-on-mmstar | LFM (w/ Qwen2.5-VL) | – | Accuracy: 66.1 |
| visual-question-answering-vqa-on-mmbench-cn | LFM (w/ Qwen2.5-VL) | – | Accuracy: 84.7 |
| visual-question-answering-vqa-on-mmbench-en | LFM (w/ Qwen2.5-VL) | – | Accuracy: 85.0 |