PDF Library Benchmarks

This benchmark is about reading pure PDF files - notscanned documents and not documents that applied OCR.

Benchmarking machine

Intel(R) Core(TM) i7 CPU 860 @ 2.80GHz

Input Documents

#	Name	File Size	Pages
1	2201.00214	2.4MiB	22
2	GeoTopo-book	5.1MiB	117
3	2201.00151	1.5MiB	12
4	1707.09725	7.0MiB	134
5	2201.00021	2.6MiB	10
6	2201.00037	2.9MiB	33
7	2201.00069	14.7MiB	15
8	2201.00178	2.3MiB	16
9	2201.00201	1.3MiB	9
10	1602.06541	2.9MiB	16
11	2201.00200	284.8KiB	7
12	2201.00022	1.2MiB	14
13	2201.00029	797.6KiB	12
14	1601.03642	1004.9KiB	8

Libraries

Name	Last PyPI Release	License	Version	Dependencies
Borb	2024-08-03	AGPL/Commercial	2.1.16
pypdfium2	2024-12-19	Apache-2.0 or BSD-3-Clause	4.30.1	PDFium (Foxit/Google)
pdfminer.six	2024-07-06	MIT/X	20231228
pdfplumber	2025-01-01	MIT	0.11.5	pdfminer.six
pdfrw	2017-09-18	MIT	0.4
pdftotext	-	GPL	0.86.1	build-essential libpoppler-cpp-dev pkg-config python3-dev
playa	2025-02-20	MIT	0.3.0
PyMuPDF	2025-02-06	GNU AFFERO GPL 3.0 / Commerical	1.25.3	MuPDF
pypdf	2025-02-09	BSD 3-Clause	5.3.0
Tika	2023-01-01	Apache v2	2.6.0	Apache Tika

Text Extraction Speed

#	Library	Average	1	2	3	4	5	6	7	8	9	10	11	12	13	14
1	pypdfium2	0.1s	0.8s	0.3s	0.2s	0.2s	0.0s	0.1s	0.1s	0.1s	0.0s	0.1s	0.0s	0.1s	0.0s	0.0s
2	PyMuPDF	0.2s	1.3s	0.4s	0.7s	0.3s	0.1s	0.2s	0.1s	0.1s	0.1s	0.1s	0.1s	0.1s	0.0s	0.0s
3	pdftotext	0.3s	1.0s	1.1s	0.3s	0.8s	0.1s	0.3s	0.2s	0.1s	0.1s	0.1s	0.1s	0.1s	0.0s	0.1s
4	playa	2.5s	17.2s	5.3s	4.4s	2.2s	0.7s	1.1s	0.6s	0.6s	0.4s	0.7s	0.5s	0.6s	0.4s	0.2s
5	pypdf	4.1s	28.7s	8.1s	8.1s	3.9s	1.2s	2.0s	0.8s	1.0s	0.8s	1.0s	0.9s	0.8s	0.6s	0.4s
6	pdfminer.six	9.0s	55.9s	23.7s	16.8s	8.9s	2.3s	4.0s	1.8s	2.2s	1.5s	2.7s	1.8s	2.0s	1.1s	0.9s
7	pdfplumber	13.0s	86.4s	22.7s	23.4s	14.2s	4.2s	7.1s	3.3s	3.2s	2.9s	4.4s	3.3s	3.5s	1.9s	1.7s
8	Tika	24.4s	17.8s	100.1s	0.6s	23.4s	47.3s	48.3s	31.5s	34.5s	0.1s	13.2s	0.1s	24.2s	0.1s	0.1s
9	Borb	50.5s	188.4s	149.1s	2.3s	113.6s	28.4s	11.7s	112.3s	23.7s	27.1s	8.4s	5.7s	27.7s	4.9s	2.9s

Image Extraction Speed

#	Library	Average	1	2	3	4	5	6	7	8	9	10	11	12	13	14
1	PyMuPDF	0.6s	0.3s	0.7s	0.0s	2.2s	0.6s	0.0s	3.3s	0.5s	0.5s	0.1s	0.0s	0.4s	0.3s	0.0s
2	pypdfium2	1.3s	1.5s	2.3s	0.0s	4.3s	1.2s	0.2s	5.7s	0.9s	0.9s	0.3s	0.1s	0.7s	0.3s	0.0s
3	pypdf	5.2s	24.6s	7.0s	6.6s	18.9s	1.7s	0.7s	7.6s	1.5s	1.5s	0.9s	0.2s	1.3s	0.3s	0.2s
4	pdfminer.six	12.3s	69.2s	24.6s	20.6s	36.6s	2.6s	4.1s	2.4s	2.3s	1.5s	2.7s	2.0s	2.1s	1.1s	0.9s

Watermarking Speed

#	Library	Average	1	2	3	4	5	6	7	8	9	10	11	12	13	14
1	pdfrw	0.1s	0.1s	0.5s	0.1s	0.4s	0.1s	0.1s	0.2s	0.1s	0.1s	0.1s	0.1s	0.2s	0.0s	0.0s
2	PyMuPDF	0.2s	0.5s	0.7s	0.2s	0.5s	0.1s	0.1s	0.1s	0.1s	0.1s	0.1s	0.0s	0.1s	0.0s	0.0s
3	pypdf	0.6s	0.7s	2.3s	0.5s	1.7s	0.3s	0.4s	0.5s	0.4s	0.2s	0.5s	0.2s	0.6s	0.1s	0.1s

Watermarking File Size

#	Library	Average	1	2	3	4	5	6	7	8	9	10	11	12	13	14
1	pypdf	3.4MB	2.5MB	5.6MB	1.6MB	7.2MB	2.7MB	3.1MB	15.4MB	2.4MB	1.3MB	3.0MB	0.3MB	1.2MB	0.8MB	1.0MB
2	pdfrw	3.5MB	2.5MB	5.7MB	1.6MB	7.3MB	2.7MB	3.1MB	15.4MB	2.4MB	1.3MB	3.0MB	0.3MB	1.2MB	0.8MB	1.0MB
3	PyMuPDF	3.7MB	2.7MB	6.9MB	1.7MB	8.5MB	2.8MB	3.4MB	15.5MB	2.5MB	1.4MB	3.2MB	0.3MB	1.3MB	0.9MB	1.1MB

Text Extraction Quality

#	Library	Average	1	2	3	4	5	6	7	8	9	10	11	12	13	14
1	pypdfium2	97%	99%	97%	94%	99%	98%	96%	99%	99%	99%	99%	98%	78%	99%	99%
2	pypdf	96%	99%	95%	93%	98%	99%	96%	97%	99%	99%	99%	99%	78%	100%	99%
3	PyMuPDF	96%	98%	96%	93%	97%	98%	95%	99%	98%	98%	98%	97%	77%	98%	99%
4	playa	96%	98%	93%	93%	98%	98%	95%	97%	97%	98%	99%	98%	77%	96%	99%
5	pdfplumber	93%	96%	89%	89%	98%	92%	94%	93%	95%	93%	97%	94%	76%	99%	98%
6	pdftotext	92%	96%	94%	91%	95%	92%	96%	96%	96%	97%	83%	94%	77%	96%	79%
7	pdfminer.six	89%	95%	79%	86%	92%	86%	93%	95%	93%	92%	92%	93%	71%	98%	86%
8	Tika	83%	99%	0%	92%	95%	77%	86%	81%	82%	98%	88%	98%	67%	98%	96%
9	Borb	45%	70%	79%	0%	40%	48%	92%	0%	64%	51%	41%	55%	41%	0%	53%