Official benchmarks map / olmOCR-bench
SoMarkAI/OXR-1.0 is ranked first on the olmOCR-bench leaderboard with 87.84, 0.3% ahead of infly/Infinity-Parser2-Pro. Read live from the Hugging Face API on November 8, 2026.
| # | Model | Organization | Score |
|---|---|---|---|
| 1 | SoMarkAI/OXR-1.0 | SoMarkAI | 87.84 |
| 2 | infly/Infinity-Parser2-Pro | infly | 87.6 |
| 3 | ProKope-AI/ProKope-4B-Eagle-Eye | ProKope-AI | 87.06 |
| 4 | lightonai/LightOnOCR-3-4B | lightonai | 86.3 |
| 5 | datalab-to/chandra-ocr-2 | datalab-to | 85.8 |
| 6 | lightonai/LightOnOCR-3-0.8B | lightonai | 85.5 |
| 7 | lightonai/LightOnOCR-3-1B | lightonai | 84.5 |
| 8 | dots-studio/dots.mocr | dots-studio | 83.9 |
| 9 | datalab-to/surya-ocr-2.5 | datalab-to | 83.7 |
| 10 | jinaai/jina-ocr-v1 | jinaai | 83.4 |
olmOCR-bench is a dataset that Hugging Face marks as an official benchmark. Its leaderboard is built automatically from the .eval_results files that model repositories publish, so entries are mostly self-reported by model authors and evaluation settings differ between submissions. A verified tag appears only where Hugging Face marks an entry as verified.
Original board: huggingface.co/datasets/allenai/olmOCR-bench
Part of the Hugging Face official benchmarks map by QuantID. Unofficial. Not affiliated with or endorsed by Hugging Face.