Document Parsing on olmOCR-bench
89.6ArXiv Processing AccuracyLightOnOCR-2-1B
Evaluation Results
| Method | Links | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| LightOnOCR-2-1BEnd-to-end=true, Size (B)=12026.01 | 89.6 | 83.2 | 89 | 42.2 | 85.6 | — | 84.8 | 91.4 | 99.6 | — | — | — | |
| LightOnOCR-2-1B-bboxEnd-to-end=true, Size (B)=12026.01 | 86.9 | 80.2 | 88.6 | 39.7 | 74.7 | — | 85 | 86.4 | 99.8 | — | — | — | |
| LightOnOCR-2-1B-ocr-soupEnd-to-end=true, Size (B)=12026.01 | 86.8 | 82.4 | 89 | 45.4 | 81.2 | — | 84.2 | 90.3 | 99.7 | — | — | — | |
| LightOnOCR-1B-1025-GRPOEnd-to-end=true, Size (B)=12026.01 | 86.5 | 77.7 | 74.5 | 32.9 | 73.8 | — | 85.1 | 91.6 | 99.7 | — | — | — | |
| LightOnOCR-2-1B-bbox-soupEnd-to-end=true, Size (B)=12026.01 | 86.1 | 80.8 | 88.2 | 41.2 | 77.9 | — | 85.4 | 87.3 | 99.7 | — | — | — | |
| PaddleOCR-VLEnd-to-end=false, Size (B)=0.92026.01 | 85.7 | 77.5 | 84.1 | 37.8 | 71 | — | 79.9 | 85.7 | 98.5 | — | — | — | |
| PaddleOCR-VL2026.01 | 85.7 | 80 | 71 | 84.1 | 37.8 | 97 | 79.9 | 85.7 | 98.5 | — | — | — | |
| Mistral OCR 3 API2026.01 | 85.6 | 79.1 | 85.5 | 43.5 | 69.7 | — | 81.2 | 88.5 | 99.7 | — | — | — | |
| LightOnOCR-2-1B-baseEnd-to-end=true, Size (B)=12026.01 | 84.9 | 81.8 | 86.7 | 47 | 80.3 | — | 84.6 | 89.1 | 99.8 | — | — | — | |
| LightOnOCR-2-1B-bbox-baseEnd-to-end=true, Size (B)=12026.01 | 84.6 | 80.8 | 84.7 | 46 | 78.6 | — | 83.8 | 88 | 99.8 | — | — | — | |
| MonkeyOCR-pro-3BEnd-to-end=false, Size (B)=32026.01 | 83.8 | 73.6 | 74.7 | 36.1 | 68.8 | — | 76.6 | 80.1 | 95.3 | — | — | — | |
| MonkeyOCR-pro-3B2026.01 | 83.8 | 75.8 | 68.8 | 74.6 | 36.1 | 91.2 | 76.6 | 80.1 | 95.3 | — | — | — | |
| Youtu-Parsing2026.01 | 83.7 | 80.5 | 71.4 | 89 | 36.7 | 94.3 | 83.1 | 87.8 | 97.8 | — | — | — | |
| olmOCR-2-8BEnd-to-end=true, Size (B)=82026.01 | 82.9 | 80.4 | 84.3 | 48.3 | 82.1 | — | 84.3 | 81.4 | 99.7 | — | — | — | |
| Chandra-9BEnd-to-end=true, Size (B)=92026.01 | 82.2 | 81.7 | 88 | 50.4 | 80.3 | — | 81.2 | 92.3 | 99.9 | — | — | — | |
| dots.ocrEnd-to-end=false, Size (B)=32026.01 | 82.1 | 76.9 | 88.3 | 40.9 | 64.2 | — | 82.4 | 81.2 | 99.5 | — | — | — | |
| dots.ocr2026.01 | 82.1 | 79.1 | 64.2 | 88.3 | 40.9 | 94.1 | 82.4 | 81.2 | 99.5 | — | — | — | |
| LightOnOCR-1B-1025End-to-end=true, Size (B)=12026.01 | 81.4 | 76.1 | 76.4 | 35.2 | 71.6 | — | 80 | 88.7 | 99.5 | — | — | — | |
| MinerU2.52026.01 | 81.1 | 77.5 | 74 | 85.1 | 33.8 | 96.3 | 65.5 | 89.8 | 94.4 | — | — | — | |
| MonkeyOCR-pro-1.2BEnd-to-end=false, Size (B)=1.22026.01 | 80.5 | 68.9 | 71.1 | 32.9 | 62.9 | — | 68.3 | 74 | 92.6 | — | — | — | |
| olmOCR v0.3.0End-to-end=false, Size (B)=82026.01 | 78.6 | 76.1 | 72.9 | 43.9 | 79.9 | — | 77.3 | 81.2 | 98.9 | — | — | — | |
| DeepSeekOCREnd-to-end=true, Size (B)=32026.01 | 77.5 | 73.1 | 77.3 | 33.1 | 74.5 | — | 67.3 | 83 | 99.3 | — | — | — | |
| MinerU2.5End-to-end=false, Size (B)=1.22026.01 | 76.6 | 70.4 | 84.9 | 33.7 | 54.6 | — | 78.2 | 81.2 | 83.5 | — | — | — | |
| MinerU2.52026.03 | 76.6 | 75.2 | 54.6 | 84.9 | 33.7 | 96.6 | 78.2 | 83.5 | 93.7 | — | — | — | |
| Marker-1.8.22026.01 | 76 | 70.1 | 57.9 | 57.6 | 27.8 | 84.9 | 72.9 | 84.6 | 99.1 | — | — | — | |
| MinerU-pipeline2026.01 | 75.4 | 61.5 | 47.4 | 60.9 | 17.3 | 96.6 | 59 | 39.1 | 96.6 | — | — | — | |
| olmOCR v0.1.75Anchor status=Anchored2026.01 | 74.9 | 75.5 | 71.2 | 71 | 42.1 | 94.5 | 78.3 | 73.3 | 98.3 | — | — | — | |
| BASELINEBackbone=DEEPSEEK-OCR-LARGE, DYNAMIC=false, #TOKENS=4002026.05 | 74.7 | 75.5 | 75.8 | 80.2 | 31.4 | 96.6 | 68.2 | 76.9 | 99.9 | — | — | — | |
| RTPruneBackbone=DEEPSEEK-OCR-LARGE, DYNAMIC=true, #TOKENS=AVG=3362026.05 | 73.7 | 73.9 | 71.8 | 80.5 | 31.7 | 95.3 | 64.6 | 73.8 | 99.5 | — | — | — | |
| RTPruneBackbone=DEEPSEEK-OCR-LARGE, DYNAMIC=false, #TOKENS=3002026.05 | 72.8 | 68.3 | 68.6 | 78.2 | 31 | 95.8 | 59 | 42.1 | 99 | — | — | — | |
| Qwen2.5-VL-72B2026.03 | 72.2 | 64.8 | 51.1 | 67.3 | 38.6 | 73.6 | 68.3 | 49.1 | 98.3 | — | — | — | |
| olmOCR v0.1.75Anchor status=No Anchor2026.01 | 71.5 | 74.7 | 71.4 | 71.4 | 42.8 | 94.1 | 77.7 | 71 | 97.8 | — | — | — | |
| BASELINEBackbone=DEEPSEEK-OCR-BASE, DYNAMIC=false, #TOKENS=2562026.05 | 69.1 | 72.4 | 76.6 | 77.7 | 30 | 95.8 | 65.3 | 65.6 | 99.3 | — | — | — | |
| MonkeyOCR-pro-3B2026.03 | 67.7 | 68.8 | 28.4 | 74.6 | 36.1 | 91.2 | 76.6 | 80.1 | 95.3 | — | — | — | |
| Nanonets-OCR-s2026.01 | 67 | 64.5 | 68.6 | 77.7 | 39.5 | 40.7 | 69.9 | 53.4 | 99.3 | — | — | — | |
| dots.ocr2026.03 | 66.3 | 73.6 | 35.8 | 88.3 | 40.9 | 94.1 | 82.4 | 81.2 | 99.5 | — | — | — | |
| RTPruneBackbone=DEEPSEEK-OCR-BASE, DYNAMIC=true, #TOKENS=AVG=2132026.05 | 65.1 | 63.7 | 69.7 | 71.1 | 29.5 | 95.3 | 45.1 | 34.8 | 99.1 | — | — | — | |
| olmOCR2026.03 | 63.9 | 71.8 | 41 | 72.9 | 43.9 | 95.1 | 77.3 | 81.2 | 98.9 | — | — | — | |
| Qwen2.5-VL-3B-PTP-0Parallel Token Prediction Strategy=02026.03 | 63.5 | 71 | 38.6 | 71.3 | 34.2 | 95.3 | 78.3 | 87.6 | 99.5 | — | — | — | |
| Qwen2.5-VL-8BEnd-to-end=true, Size (B)=82026.01 | 63.1 | 64.3 | 67.3 | 38.6 | 65.7 | — | 68.3 | 49.1 | 98.3 | — | — | — | |
| Qwen2.5-VL-7BAnchor status=No Anchor2026.01 | 63.1 | 65.5 | 65.7 | 67.3 | 38.6 | 73.6 | 68.3 | 49.1 | 98.3 | — | — | — | |
| Qwen2.5-VL-3B-PTP-1Parallel Token Prediction Strategy=12026.03 | 62.6 | 70.6 | 38.6 | 70.4 | 33.6 | 96.2 | 79.4 | 84.7 | 99.4 | — | — | — | |
| MinerU2-pipeline2026.03 | 61.8 | 55.6 | 13.5 | 60.9 | 17.3 | 96.6 | 59 | 39.1 | 96.6 | — | — | — | |
| RTPruneBackbone=DEEPSEEK-OCR-BASE, DYNAMIC=false, #TOKENS=1922026.05 | 59.5 | 57.7 | 65.3 | 62.8 | 28.5 | 94.3 | 31.8 | 20.4 | 98.6 | — | — | — | |
| VISIONZIPBackbone=DEEPSEEK-OCR-LARGE, DYNAMIC=true, #TOKENS=AVG=3362026.05 | 58 | 59.7 | 60.9 | 63.4 | 30.8 | 96.4 | 27.7 | 41.2 | 99 | — | — | — | |
| Gemini Flash 2Anchor status=Anchored2026.01 | 54.5 | 63.8 | 56.1 | 72.1 | 34.2 | 64.7 | 61.5 | 71.5 | 95.6 | — | — | — | |
| GPT-4oAnchor status=Anchored2026.01 | 53.5 | 69.9 | 74.5 | 70 | 40.7 | 93.8 | 69.3 | 60.6 | 96.8 | — | — | — | |
| GOT2026.01 | 52.7 | 48.3 | 52 | 0.2 | 22.1 | 93.6 | 42 | 29.9 | 94 | — | — | — | |
| GPT-4oAnchor status=No Anchor2026.01 | 51.5 | 68.9 | 75.5 | 69.1 | 40.9 | 94.2 | 68.9 | 54.1 | 96.7 | — | — | — | |
| CDPRUNERBackbone=DEEPSEEK-OCR-BASE, DYNAMIC=true, #TOKENS=AVG=2132026.05 | 47.2 | 53 | 56.5 | 65.5 | 24.5 | 95.1 | 15 | 21.5 | 98.9 | — | — | — | |
| GPT-4o2026.03 | 44.1 | 63.2 | 37.6 | 69.1 | 40.9 | 94.2 | 68.9 | 54.1 | 96.7 | — | — | — | |
| DIVPRUNEBackbone=DEEPSEEK-OCR-BASE, DYNAMIC=true, #TOKENS=AVG=2132026.05 | 42.2 | 50.1 | 42.8 | 48 | 25.1 | 95.7 | 22.3 | 25.8 | 98.9 | — | — | — | |
| Gemini Flash 22026.01 | 32.1 | 59.2 | 61.4 | 27.8 | 56.3 | — | 58.7 | 84.4 | 94 | — | — | — | |
| Gemini Flash 2Anchor status=No Anchor2026.01 | 32.1 | 57.8 | 56.3 | 61.4 | 27.8 | 48 | 58.7 | 84.4 | 94 | — | — | — | |
| SPARSEVLMBackbone=DEEPSEEK-OCR-LARGE, DYNAMIC=true, #TOKENS=AVG=3362026.05 | 16.1 | 32 | 6.3 | 12.5 | 16.3 | 93.6 | 6.2 | 17.6 | 87.3 | — | — | — | |
| FITPRUNEBackbone=DEEPSEEK-OCR-BASE, DYNAMIC=false, #TOKENS=1922026.05 | 12.9 | 34.1 | 25.8 | 27.2 | 17.1 | 92.6 | 5.1 | 4.8 | 87.6 | — | — | — | |
| FASTVBackbone=DEEPSEEK-OCR-LARGE, DYNAMIC=false, #TOKENS=3002026.05 | 1.4 | 25.2 | 3.1 | 2.8 | 15 | 97.1 | 0.1 | 0.2 | 82.1 | — | — | — | |
| DARTBackbone=DEEPSEEK-OCR-LARGE, DYNAMIC=false, #TOKENS=3002026.05 | 1.2 | 25.6 | 3.4 | 3 | 15.4 | 98.6 | 0.1 | 0.1 | 84.5 | — | — | — | |
| NUWABackbone=DEEPSEEK-OCR-BASE, DYNAMIC=false, #TOKENS=1922026.05 | 0.2 | 23.4 | 2.2 | 2.7 | 14.1 | 97.1 | 0 | 0 | 70.7 | — | — | — |