ocr-bench
Leaderboard
Comparisons
Winner
All
A
B
tie
Model
All
FireRedTeam/FireRed-OCR
PaddlePaddle/PP-OCRv6_medium
PaddlePaddle/PaddleOCR-VL-1.6
allenai/olmOCR-2-7B-1025-FP8
baidu/Qianfan-OCR
baidu/Unlimited-OCR
deepseek-ai/DeepSeek-OCR
deepseek-ai/DeepSeek-OCR-2
lightonai/LightOnOCR-2-1B
numind/NuExtract3
rednote-hilab/dots.mocr
rednote-hilab/dots.ocr
tesseract-5
zai-org/GLM-OCR
Model A Model B W L T FireRed-OCR PP-OCRv6_medium 32 16 0 FireRed-OCR PaddleOCR-VL-1.6 25 23 1 FireRed-OCR olmOCR-2-7B-1025-FP8 20 30 0 FireRed-OCR Qianfan-OCR 25 24 1 FireRed-OCR Unlimited-OCR 23 23 0 FireRed-OCR DeepSeek-OCR 31 19 0 FireRed-OCR DeepSeek-OCR-2 37 13 0 FireRed-OCR LightOnOCR-2-1B 15 29 0 FireRed-OCR NuExtract3 19 31 0 FireRed-OCR dots.mocr 14 31 0 FireRed-OCR dots.ocr 49 1 0 FireRed-OCR tesseract-5 37 9 0 FireRed-OCR GLM-OCR 11 37 2 PP-OCRv6_medium PaddleOCR-VL-1.6 15 33 0 PP-OCRv6_medium olmOCR-2-7B-1025-FP8 12 36 0 PP-OCRv6_medium Qianfan-OCR 15 33 0 PP-OCRv6_medium Unlimited-OCR 15 31 0 PP-OCRv6_medium DeepSeek-OCR 12 36 0 PP-OCRv6_medium DeepSeek-OCR-2 20 28 0 PP-OCRv6_medium LightOnOCR-2-1B 5 39 0 PP-OCRv6_medium NuExtract3 6 42 0 PP-OCRv6_medium dots.mocr 6 39 0 PP-OCRv6_medium dots.ocr 48 0 0 PP-OCRv6_medium tesseract-5 45 1 0 PP-OCRv6_medium GLM-OCR 5 43 0 PaddleOCR-VL-1.6 olmOCR-2-7B-1025-FP8 13 36 0 PaddleOCR-VL-1.6 Qianfan-OCR 10 39 0 PaddleOCR-VL-1.6 Unlimited-OCR 12 34 0 PaddleOCR-VL-1.6 DeepSeek-OCR 31 18 0 PaddleOCR-VL-1.6 DeepSeek-OCR-2 28 21 0 PaddleOCR-VL-1.6 LightOnOCR-2-1B 4 40 0 PaddleOCR-VL-1.6 NuExtract3 14 35 0 PaddleOCR-VL-1.6 dots.mocr 1 44 0 PaddleOCR-VL-1.6 dots.ocr 49 0 0 PaddleOCR-VL-1.6 tesseract-5 43 3 0 PaddleOCR-VL-1.6 GLM-OCR 6 43 0 olmOCR-2-7B-1025-FP8 Qianfan-OCR 33 17 0 olmOCR-2-7B-1025-FP8 Unlimited-OCR 38 8 0 olmOCR-2-7B-1025-FP8 DeepSeek-OCR 40 9 1 olmOCR-2-7B-1025-FP8 DeepSeek-OCR-2 44 5 1 olmOCR-2-7B-1025-FP8 LightOnOCR-2-1B 16 26 1 olmOCR-2-7B-1025-FP8 NuExtract3 30 19 0 olmOCR-2-7B-1025-FP8 dots.mocr 30 13 1 olmOCR-2-7B-1025-FP8 dots.ocr 50 0 0 olmOCR-2-7B-1025-FP8 tesseract-5 46 0 0 olmOCR-2-7B-1025-FP8 GLM-OCR 25 25 0 Qianfan-OCR Unlimited-OCR 30 16 0 Qianfan-OCR DeepSeek-OCR 30 20 0 Qianfan-OCR DeepSeek-OCR-2 37 12 1 Qianfan-OCR LightOnOCR-2-1B 13 31 0 Qianfan-OCR NuExtract3 18 32 0 Qianfan-OCR dots.mocr 4 41 0 Qianfan-OCR dots.ocr 48 2 0 Qianfan-OCR tesseract-5 44 2 0 Qianfan-OCR GLM-OCR 10 40 0 Unlimited-OCR DeepSeek-OCR 32 14 0 Unlimited-OCR DeepSeek-OCR-2 37 9 0 Unlimited-OCR LightOnOCR-2-1B 12 32 0 Unlimited-OCR NuExtract3 11 35 0 Unlimited-OCR dots.mocr 6 38 0 Unlimited-OCR dots.ocr 46 0 0 Unlimited-OCR tesseract-5 44 1 0 Unlimited-OCR GLM-OCR 12 34 0 DeepSeek-OCR DeepSeek-OCR-2 30 18 1 DeepSeek-OCR LightOnOCR-2-1B 4 39 1 DeepSeek-OCR NuExtract3 8 42 0 DeepSeek-OCR dots.mocr 7 37 1 DeepSeek-OCR dots.ocr 46 4 0 DeepSeek-OCR tesseract-5 42 4 0 DeepSeek-OCR GLM-OCR 7 43 0 DeepSeek-OCR-2 LightOnOCR-2-1B 4 39 1 DeepSeek-OCR-2 NuExtract3 14 34 0 DeepSeek-OCR-2 dots.mocr 5 39 1 DeepSeek-OCR-2 dots.ocr 45 5 0 DeepSeek-OCR-2 tesseract-5 31 15 0 DeepSeek-OCR-2 GLM-OCR 6 44 0 LightOnOCR-2-1B NuExtract3 23 21 0 LightOnOCR-2-1B dots.mocr 23 20 1 LightOnOCR-2-1B dots.ocr 43 1 0 LightOnOCR-2-1B tesseract-5 44 0 0 LightOnOCR-2-1B GLM-OCR 18 26 0 NuExtract3 dots.mocr 23 22 0 NuExtract3 dots.ocr 46 4 0 NuExtract3 tesseract-5 43 3 0 NuExtract3 GLM-OCR 32 18 0 dots.mocr dots.ocr 45 0 0 dots.mocr tesseract-5 44 0 0 dots.mocr GLM-OCR 23 22 0 dots.ocr tesseract-5 5 41 0 dots.ocr GLM-OCR 2 48 0 tesseract-5 GLM-OCR 0 46 0
A
The image is too blurry to recognize any text content.
B
The provided image is a photograph of a textured, brown surface (likely a piece of aged paper or a painting) and does not contain any text, mathematical formulas, or tables to convert.
Which OCR output is better?
1 vote
·
0% agree
· judge may be miscalibrated