Scenario in benchmark
A cleanly scanned document
Its role in Converting a complex PDF into clean Markdown with a hosted API.
Read the neutral global definition →Tools on this scenario
3 of 13 tools have a published result. Open a result to see what was tested and its evidence.
| Tool | Published result | Result and evidence |
|---|---|---|
| Adobe PDF Extract API | No published result | |
| Extend AI | 1 of 1 test case passed | Can Extend AI read a cleanly scanned document accurately? → Version 1 · Round 1 · Graded 29 Sept 2026 |
| Firecrawl | No published result | |
| GPT-5.6 terra | No published result | |
| Landing AI | No published result | |
| LlamaParse | No published result | |
| Mistral AI | 1 of 1 test case passed | Can Mistral AI read a cleanly scanned document? → Version 1 · Round 1 · Graded 5 Oct 2026 |
| Nutrient.io | No published result | |
| PDF Vector | No published result | |
| Reducto | 1 of 1 test case passed | Can Reducto read a cleanly scanned document accurately? → Version 1 · Round 1 · Graded 29 Sept 2026 |
| Tensorlake | No published result | |
| Unstructured | No published result | |
| Upstage AI | No published result | |
How this benchmark uses it
- Role
- context
- Pinned test cases
- 1
- Disclosure
- 1 public · 0 withheld