BENCHMARKS Open benchmarks, every parser we could run, every test public. ← All benchmarks

TABLES · OMNIPARSEBENCH · TABLE TAG

Tables: Datalab accurate scores 95.6% on OmniParseBench · table tag.

Every OmniParseBench test that touches a table: 6,767 tests from 875 documents. A table_cell test passes only when a cell sits under the right column and row headings and beside the right neighbors, so merged cells, spanning headers and split tables all count.

Try it free Talk to sales →

SCOREBOARD · OmniParseBench · table tag

Ranked scoreboard.

Models ranked highest-to-lowest. Datalab variants in accent; competitors and prior generations in muted ink.

Rank Model Score vs scale

01 Datalab accurate production run, Oct 23 95.6%

02 Extend 94.7%

03 Datalab balanced production run, Oct 23 94.2%

04 Gemini 3.8 Flash* 93.8%

05 LlamaParse 93.7%

06 Reducto 92.7%

07 Claude Sonnet 5.5* 92.0%

08 GPT-5.6 Sol 91.6%

09 Mistral OCR 91.5%

10 Azure 86.1%

#1 of 10 parsers

6,767 tests · 875 documents

* Used in creating and verifying at least some OmniParseBench tests.

WHY THIS NUMBER IS CREDIBLE

Every OmniParseBench test is a yes/no question about one page, and every score is the share of tests passed. The scorer is open code with no model judging the outputs. Each tag score here is the mean of the text, tables and layout pass rates on the tests with that tag. Score any slice yourself in the OmniParseBench post, or download the tests from the Hugging Face dataset.

Read the full results →

OTHER BENCHMARKS

Compare on another doc type.

All benchmarks →

Multi-column Newspapers · journals · filings 94.9% 94.9 for Datalab accurate, the top score of every parser. Reading order across columns, spreads and sidebars. Overall Every test: text, tables and layout 94.1% 94.1 for Datalab accurate, the top score of every parser we ran.

START

Nested headers, merged cells, colspans.

Run your hardest table through Datalab. Free tier, no credit card.

Try in Playground Sign up →