LlamaParse in Converting a complex PDF into clean Markdown with a hosted API
Scenario-level performance from current published Results.
10 of 12 scenarios published
Where LlamaParse failed
2 of 10 graded scenarios. Open a Result to see the conversation and the evidence.
A document with figures and captions
LlamaParse replaced the figure with prose rather than preserving it in the default Markdown output. Its caption remained in place, but the image was neither embedded nor linked.
View Result →A document with styled headings and subheadings
LlamaParse turned every heading into a Markdown heading but kept only two of the source's three heading levels. On page 1, the subsections Chain of custody and Condition on arrival came out at the same level as their parent, Reception, and the third-tier headings shared a level with the second. A separate run with a custom prompt asking for three heading levels did not preserve the hierarchy either.
View Result →How LlamaParse performed
Open a capability to explore its scenarios. Each row reports the test set in its published Result.
Code Extraction1 pass
Capability in this benchmark →
| Scenario | Published outcomes | Test coverage | Result |
|---|---|---|---|
| A document containing a code block | 1 Pass0 Fail0 Not gradable 1 of 1 test case passedWhat happenedLlamaParse returned both code listings in the document as fenced blocks tagged as Python and shell, and kept the Python body at its three indent levels of 4, 8 and 12 spaces. Two small whitespace differences remain inside the fences: one blank line was dropped, and the shell continuation line sits at four spaces rather than six. | 1/1 assessed1/1 gradablePublished test set | View Result → |
Equations & Mathematical Notation1 pass
Capability in this benchmark →
| Scenario | Published outcomes | Test coverage | Result |
|---|---|---|---|
| A document containing mathematical equations | 1 Pass0 Fail0 Not gradable 1 of 1 test case passedWhat happenedLlamaParse returned all six numbered equations of a three-page technical note as LaTeX math, each matching the printed page symbol by symbol, and none came out as garbled prose. Inline math stayed inline and Greek letters came back as LaTeX commands. | 1/1 assessed1/1 gradablePublished test set | View Result → |
Figures & Charts1 fail1 no Result yet
Capability in this benchmark →
| Scenario | Published outcomes | Test coverage | Result |
|---|---|---|---|
| A document with figures and captions | 0 Pass1 Fail0 Not gradable 1 of 1 test case failedWhat happenedLlamaParse replaced the figure with prose rather than preserving it in the default Markdown output. Its caption remained in place, but the image was neither embedded nor linked. | 1/1 assessed1/1 gradablePublished test set | View Result → |
| A document with a data chart | No published result | ||
Heading & Section Structure1 pass1 fail
Capability in this benchmark →
| Scenario | Published outcomes | Test coverage | Result |
|---|---|---|---|
| A document with styled headings and subheadings | 0 Pass1 Fail0 Not gradable 1 of 1 test case failedWhat happenedLlamaParse turned every heading into a Markdown heading but kept only two of the source's three heading levels. On page 1, the subsections Chain of custody and Condition on arrival came out at the same level as their parent, Reception, and the third-tier headings shared a level with the second. A separate run with a custom prompt asking for three heading levels did not preserve the hierarchy either. | 1/1 assessed1/1 gradablePublished test set | View Result → |
| Footnotes at the bottom of the page | 1 Pass0 Fail0 Not gradable 1 of 1 test case passedWhat happenedLlamaParse kept all 13 footnotes out of the body text and tied each one to its reference. Each page's notes come after that page's body, every note on its own line, and every reference mark in the body is a superscript that matches a note, with every note matching a mark. | 1/1 assessed1/1 gradablePublished test set | View Result → |
Reading Order & Layout1 pass
Capability in this benchmark →
| Scenario | Published outcomes | Test coverage | Result |
|---|---|---|---|
| A page laid out in multiple columns | 1 Pass0 Fail0 Not gradable 1 of 1 test case passedWhat happenedLlamaParse read the two-column bulletin in natural order on all three pages. It finished each column before continuing to the next, without interleaving the text. | 1/1 assessed1/1 gradablePublished test set | View Result → |
Scanned Document OCR2 pass
Capability in this benchmark →
| Scenario | Published outcomes | Test coverage | Result |
|---|---|---|---|
| A cleanly scanned document | 1 Pass0 Fail0 Not gradable 1 of 1 test case passedWhat happenedLlamaParse recovered every printed word of the image-only certificate correctly, including the station code VQ1, all five mg/l readings, the four method codes and the dates. The six-row results block came back as a table, and the only addition was a line standing in for the handwritten signature. | 1/1 assessed1/1 gradablePublished test set | View Result → |
| A document mixing digital and scanned pages | 1 Pass0 Fail0 Not gradable 1 of 1 test case passedWhat happenedLlamaParse processed the scanned page in a PDF that also contained digital pages. Its Markdown included the scanned worksheet’s text, readings, identifiers and times. | 1/1 assessed1/1 gradablePublished test set | View Result → |
Table Extraction1 pass1 no Result yet
Capability in this benchmark →
| Scenario | Published outcomes | Test coverage | Result |
|---|---|---|---|
| A simple, clearly formatted table | 1 Pass0 Fail0 Not gradable 1 of 1 test case passedWhat happenedLlamaParse returned the page-1 charges table intact: the four headers in order, all eight rows and every value under the right column. The cells most likely to drift held: 'Metals suite, eleven elements' stayed one cell, and '10 days' stayed out of the charge column. The table came back as an HTML table inside the Markdown. | 1/1 assessed1/1 gradablePublished test set | View Result → |
| A table that continues across a page break | No published result | ||
Text Fidelity1 pass
Capability in this benchmark →
| Scenario | Published outcomes | Test coverage | Result |
|---|---|---|---|
| An ordinary digital text document | 1 Pass0 Fail0 Not gradable 1 of 1 test case passedWhat happenedLlamaParse preserved the briefing note’s body text word for word: 3,122 words in the source and 3,122 in Markdown. All eleven money and measurement figures were present. | 1/1 assessed1/1 gradablePublished test set | View Result → |
Reading these Results
Published evidence and test coverage answer different questions.
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.
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.
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 LlamaParse was tested on it.