| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| OCRBench | Routing | Score922 | 232 | 3d ago | |
| OCRVQA (test) | OCRVQA Score74.87 | 38 | 1mo ago | ||
| OCRBench (test) | InternVL2-2B | Score78.1 | 34 | 1mo ago | |
| ORB-En-SynthDog 4-class rotation | Nanonets-OCR-s | Word Error Rate (WER)2.61 | 27 | 1mo ago | |
| ORB-En-FUNSD (test) | Nanonets-OCR-s | Word ANLS99.58 | 24 | 1mo ago | |
| UPTI (val) | UTRNet | Accuracy95.97 | 22 | 1mo ago | |
| IIITH (val) | UTRNet | Accuracy88.01 | 22 | 1mo ago | |
| UTRSet-Real (val) | UTRNet | Accuracy92.97 | 22 | 1mo ago | |
| ORB-Indic (test) | GPT-4o | Word ANLS97.7 | 21 | 1mo ago | |
| ORB-Indic 4-class rotation | GPT-4o | Word Accuracy94.55 | 21 | 1mo ago | |
| OCRBench English v2 | Recognition73.2 | 19 | 1mo ago | ||
| Chronicle American LoCCA (eval) | Character Error Rate0.5 | 18 | 1mo ago | ||
| Dongba Script (test) | OmniOCR | Accuracy95.32 | 16 | 1mo ago | |
| Shui Script (test) | OmniOCR | Accuracy95.95 | 16 | 1mo ago | |
| Ancient Yi (test) | Accuracy90.53 | 16 | 1mo ago | ||
| Tibetan (test) | OmniOCR | Accuracy90.37 | 16 | 1mo ago | |
| Ancient Greek | TrOCR (Base) | Character Error Rate2.7 | 15 | 1mo ago | |
| OCRBench v2 | Accuracy (en)68 | 14 | 5d ago | ||
| TextOCR VL-ICL Bench | LLaVA-OneVision-72B | Average Accuracy52.2 | 14 | 1mo ago | |
| Japanese Horizontal | EffOCR-C (Base) | Character Error Rate0.6 | 12 | 1mo ago | |
| OCRBench Text | Pure VLM Score95 | 10 | 1mo ago | ||
| olmOCR multiple category 1.5 | Average Accuracy82 | 9 | 18d ago | ||
| CCOCR | Qianfan-OCR | Multi-language Score76.7 | 9 | 1mo ago | |
| WuDao rendered text images (test) | Deepseek-OCR | ROUGE (R=2)0.996 | 9 | 1mo ago | |
| TextOCR | Accuracy92.5 | 9 | 1mo ago |
