Does PDF Vector keep a figure with its caption in place?
PDF Vector failed to keep the figure with its caption in place. The caption stayed in the right spot, but the figure itself was missing from the Markdown and the response, and the run card saved no asset files.
1 of 1 test case failed
Every test case in this scenario has a result.
The test case
Each test case is judged on its own: Pass, Fail, Not gradable, or Untested. The scenario result above is based on this test case.
Figure with caption — preserved in placeFailEvidence
We sent the complete four-page survey report PDF to PDF Vector's document parse endpoint in one request; there was no typed prompt.
Markdown output with the caption in place and no figure
Copied from Output file (Markdown), lines 49–50
Figure 1. Survey transects at Kellow Sands and Thrimby Point. Sampling stations are shown as filled circles and the 1998 saltmarsh edge as a broken line.
Raw API response with only text output and no image fields
Copied from Output file (JSON) · an excerpt
{ "markdown": "Intertidal survey of Kellow Sands and Thrimby Point\nReport BEP/SR/2026/11 · Survey of 19 to 23 January 2026 · Issued February 2026\nSummary\nThe four fixed transects at Kellow Sands and Thrimby Point were resurveyed in January, in the third w … [5,688 characters in the file]", "pageCount": 4, "model": "nano", "credits": 4, "requestId": 272368 }
Excerpt ends here. Open the source for the complete output.
Open the whole JSON file ↗| ✗ | Expected: The image is preserved at its position, with its caption.Found: The caption stays in place, but the figure is missing from the Markdown and raw response. |
The default full-PDF run succeeded, but the figure never appeared while the caption did, so this is a product fail rather than a run problem.
Figure labels not returned either. Under the default pipeline the plan's own labels, such as Kellow Sands, Thrimby Point and the transect names T1 to T4, were not returned as text either, so nothing of the figure survives in the output.
Configuration and setup
How this tool was set up for the run and what the test needed in place. Each row is a fact from the run's records; a fact the records do not hold is left out, not guessed.
How this scenario is graded
How we decide Pass, Fail and Not gradable. The same rules apply to every tool tested on this scenario.
Each test case gets one result per tool: Pass, Fail or Not gradable. A test case we haven't run yet shows Untested. There are no partial results.
- Pass: the tool did everything the test expected, and nothing it said contradicts the correct answer.
- Fail: at least one expected behaviour clearly didn't happen; the test case's result says which one and quotes the tool.
- Not gradable: our evidence couldn't settle the outcome (for example a record we needed is missing). It is never counted as a fail, and the test case's result says what's missing.
Where this sits in the benchmark
This page is one cell of a larger study: one tool, one scenario. Only this benchmark's frame appears here.
| Level | Name | Scope |
|---|---|---|
| Benchmark | Converting a complex PDF into clean Markdown with a hosted API → | 12 scenarios · 13 tools |
| Capability | Figures & Charts → | |
| Scenario | A document with figures and captions → | |
| Tool | PDF Vector → |
Global scenario definition → · Global capability definition → · PDF Vector product page →
History of this result
What has happened to this result since it was first published. Runs and grades are never overwritten: a retest or a re-grade publishes a new result and keeps the earlier one readable.
Act on this result
Nothing filed here edits the run or the grade. A challenge opens a review, and a review can produce a new run or a re-grade — which becomes the current result and leaves this one in the history.
You run the same kind of test against your own setup and get the same behaviour.
Agree →Yours behaves differently. Tell us what you got, with a screenshot if you have one.
Disagree →Something here is wrong — a reference value, a transcription, a grade.
Report an issue →Tell us what changed and we schedule a rerun of the failing test case. The old result stays as history.
Vendor notice →The same record is available as structured data through the AI Demos MCP server, with the counts, the coverage and every per-test-case reason carried as fields.
Verify the proof files
These files support this result. Open a file to inspect the original evidence.
File fingerprints (SHA-256)
A fingerprint identifies the exact file used for this result.