Tool in benchmark · Version 1

Docling in Converting a complex PDF into clean Markdown with an open-source library

Scenario-level performance from current published Results.

4Passed
6Failed
2No published Result yet

10 of 12 scenarios published

CapabilityScenariosResult
Code Extraction0 pass · 1 fail · 1 of 1 published
Equations & Mathematical Notation1 pass · 0 fail · 1 of 1 published
Figures & Charts0 pass · 2 fail · 2 of 2 published
Heading & Section Structure0 pass · 2 fail · 2 of 2 published
Reading Order & Layout1 pass · 0 fail · 1 of 1 published
Scanned Document OCR1 pass · 0 fail · 1 of 2 published
Table Extraction1 pass · 0 fail · 1 of 2 published
Text Fidelity0 pass · 1 fail · 1 of 1 published
PassedFailedPublished, different test setNo published Result yet

Where Docling failed

6 of 10 graded scenarios. Open a Result to see the conversation and the evidence.

Code Extraction

A document containing a code block

Docling failed to keep code block line structure and indentation intact. It fenced the Python, but the code and shell blocks collapsed to single lines, and the block-header excerpt became ordinary prose.

View Result →
Figures & Charts

A document with a data chart

Docling’s default Markdown conversion did not preserve the data chart as an image. The chart title became a heading, followed by an image placeholder at the correct position.

View Result →
Figures & Charts

A document with figures and captions

Docling’s default Markdown export omitted the figure images. It kept the transect drawing’s caption and the bar chart’s title in their reading positions, with image placeholders rather than embedded pictures or image references.

View Result →
Heading & Section Structure

A document with styled headings and subheadings

Docling did not keep the heading hierarchy intact for this PDF. The source showed three heading levels and a title block, but the Markdown flattened all 26 headings to level 2 and merged one section with its subsection.

View Result →
Heading & Section Structure

Footnotes at the bottom of the page

Docling did not preserve the relationship between footnotes and their references in Markdown. It included all 13 footnote texts, but placed them in the body flow without associating them with their markers.

View Result →
Text Fidelity

An ordinary digital text document

Docling omitted four body words when converting the digital document to Markdown. The next paragraph was joined to an unfinished sentence; nothing was garbled, invented or duplicated.

View Result →

How Docling performed

Open a capability to explore its scenarios. Each row reports the test set in its published Result.

Code Extraction1 fail
ScenarioPublished outcomesTest coverageResult
A document containing a code block
0 Pass1 Fail0 Not gradable
1 of 1 test case failed
What happened

Docling failed to keep code block line structure and indentation intact. It fenced the Python, but the code and shell blocks collapsed to single lines, and the block-header excerpt became ordinary prose.

1/1 assessed1/1 gradablePublished test setView Result →
Equations & Mathematical Notation1 pass
ScenarioPublished outcomesTest coverageResult
A document containing mathematical equations
1 Pass0 Fail0 Not gradable
1 of 1 test case passed
What happened

Docling passed this equation-handling case. It converted the three-page PDF in the default pipeline and replaced all six display equations with explicit fallback markers in the right places instead of garbled prose. The inline mathematical symbols were moved out of some sentences, and no LaTeX or MathML appears in the output.

1/1 assessed1/1 gradablePublished test setView Result →
Figures & Charts2 fail
ScenarioPublished outcomesTest coverageResult
A document with a data chart
0 Pass1 Fail0 Not gradable
1 of 1 test case failed
What happened

Docling’s default Markdown conversion did not preserve the data chart as an image. The chart title became a heading, followed by an image placeholder at the correct position.

1/1 assessed1/1 gradablePublished test setView Result →
A document with figures and captions
0 Pass1 Fail0 Not gradable
1 of 1 test case failed
What happened

Docling’s default Markdown export omitted the figure images. It kept the transect drawing’s caption and the bar chart’s title in their reading positions, with image placeholders rather than embedded pictures or image references.

1/1 assessed1/1 gradablePublished test setView Result →
Heading & Section Structure2 fail
ScenarioPublished outcomesTest coverageResult
A document with styled headings and subheadings
0 Pass1 Fail0 Not gradable
1 of 1 test case failed
What happened

Docling did not keep the heading hierarchy intact for this PDF. The source showed three heading levels and a title block, but the Markdown flattened all 26 headings to level 2 and merged one section with its subsection.

1/1 assessed1/1 gradablePublished test setView Result →
Footnotes at the bottom of the page
0 Pass1 Fail0 Not gradable
1 of 1 test case failed
What happened

Docling did not preserve the relationship between footnotes and their references in Markdown. It included all 13 footnote texts, but placed them in the body flow without associating them with their markers.

1/1 assessed1/1 gradablePublished test setView Result →
Reading Order & Layout1 pass
ScenarioPublished outcomesTest coverageResult
A page laid out in multiple columns
1 Pass0 Fail0 Not gradable
1 of 1 test case passed
What happened

Docling preserved the natural reading order in a multi-column document. Across three two-column pages, it read the left column before the right without mixing their text, and kept the sentences spanning column breaks together.

1/1 assessed1/1 gradablePublished test setView Result →
Scanned Document OCR1 pass1 no Result yet
ScenarioPublished outcomesTest coverageResult
A document mixing digital and scanned pages
1 Pass0 Fail0 Not gradable
1 of 1 test case passed
What happened

Docling included the scanned worksheet's text and values in the Markdown output instead of skipping the page. The worksheet stayed in its original position between the surrounding pages.

1/1 assessed1/1 gradablePublished test setView Result →
A cleanly scanned documentNo published result
Table Extraction1 pass1 no Result yet
ScenarioPublished outcomesTest coverageResult
A simple, clearly formatted table
1 Pass0 Fail0 Not gradable
1 of 1 test case passed
What happened

Docling kept the simple table intact in Markdown. The output has one four-column table with eight data rows, and every value stayed under the right header.

1/1 assessed1/1 gradablePublished test setView Result →
A table that continues across a page breakNo published result
Text Fidelity1 fail
ScenarioPublished outcomesTest coverageResult
An ordinary digital text document
0 Pass1 Fail0 Not gradable
1 of 1 test case failed
What happened

Docling omitted four body words when converting the digital document to Markdown. The next paragraph was joined to an unfinished sentence; nothing was garbled, invented or duplicated.

1/1 assessed1/1 gradablePublished test setView Result →

Reading these Results

Published evidence and test coverage answer different questions.

Publication availability

Which scenarios have a Result?

A published Result is public evidence for this tool on one scenario. “No published result” does not say whether testing has taken place.

Test coverage

What does each Result cover?

Assessed includes Pass, Fail and Not gradable. Gradable includes Pass and Fail. Both use the pinned test count in that published Result.

Scenario scope

Out of 12 scenarios

The summary counts each of the 12 scenarios once. A scenario fails when any of its published test cases failed. “No published Result yet” means nothing is public for that scenario; it does not say whether Docling was tested on it.

Docling in Converting a complex PDF into clean Markdown with an open-source library | AI Demos