Omni Parse Bench

A document-parsing benchmark. Every test is a yes/no question about one page: is this text in the output, in this order, in a table cell under these headings, in a layout block of this kind?

This repository contains a way to run our benchmark using our dataset on HuggingFace or your own dataset. It also provides lower-level primitives (score and predict) to run in your own pipelines however you like. The README is split up into the following table of contents.

• The design
• Install
• Run the benchmark
• Score
• Predict
• Documentation
• License
• Citation

The design

The headline is the mean of three family scores: text, tables and layout. A family is a group of test types:

A test checks one or more args, each of which is the content at one place on the page. Each arg is tagged:

• structure: table, math, form and multi_column;
• rendering: handwriting, tiny, rotated and degraded;
• role: heading, caption, footnote, list, code, figure and header_footer;
• script: ISO 15924 codes (Latn, Hani, Deva, ...); and
• language: ISO 639-3 codes (eng, fra, tha, ...).

A test's tags are the union of its args' tags. Tags never decide whether a test passes or what the headline is; they select tests, so any slice of the benchmark can be scored. A test with several tags counts in each, so per-tag scores overlap and don't average to the headline.

Install

uv pip install omni-parse-bench                 # score(): no network, no API keys
uv pip install 'omni-parse-bench[harness]'      # + vendor adapters, to predict
uv pip install 'omni-parse-bench[benchmark]'    # + dataset access, to run the benchmark

Run the benchmark

With [benchmark] installed, run vendors on the dataset (5 pages here):

opb benchmark --out runs/ --limit 5 --providers datalab openai/gpt-5.6-sol

It shows the plan and asks before calling any vendor, since calls cost money. Each vendor needs its API key in the environment: DATALAB_API_KEY for Datalab, OPENROUTER_API_KEY for any model by its OpenRouter id; docs/providers.md lists the rest. A run is resumable: stop it, and run the same command again to carry on.

Each setting of a provider is its own run; here, two:

opb benchmark --out runs/ --limit 5 --providers datalab \
  --options '{"datalab": [{"mode": "balanced"}, {"mode": "accurate"}]}'

--dataset picks the dataset: a folder, or hf://datasets/<repo>[@revision] (ours by default). --suites, --sample and --pages choose pages. docs/API.md has every option, how resuming works and what a run writes. Each run's scores.jsonl has one line per test; Asking a question shows how to compare runs by tag, suite or test type.

Score

Here is one page with three tests, written inline with gold.parse so the gold file's shape is visible (gold.read("gold/<page_id>.json") reads a file):

from omni_parse_bench import gold, score, summarize

tests = gold.parse("example-p1", {"tests": [
    {
        "id": "t1",
        "test_type": "present",
        "output_type": "html",
        "derivation": "text_layer",
        "tags": ["Latn", "eng"],
        "args": {"content": {"text": "Quarterly revenue rose 12%", "max_diffs": 1}},
        "settings": {},
    },
    {
        "id": "t2",
        "test_type": "order",
        "output_type": "html",
        "derivation": "text_layer",
        "tags": ["Latn", "eng", "heading"],
        "args": {
            "before": {"text": "Summary", "max_diffs": 0},
            "after": {"text": "Outlook", "max_diffs": 0},
        },
        "settings": {},
    },
    {
        "id": "t3",
        "test_type": "table_cell",
        "output_type": "html",
        "derivation": "publisher_markup",
        "tags": ["table"],
        "args": {
            "cell": {"text": "4.2", "max_diffs": 0},
            "top_heading": {"text": "Q3", "max_diffs": 0},
        },
        "settings": {},
    },
]})

outputs = {
    "html": (
        "<h2>Summary</h2>"
        "<p>Quarterly revenue rose 12% on the year.</p>"
        "<table>"
        "<tr><th>Q2</th><th>Q3</th></tr>"
        "<tr><td>3.9</td><td>4.1</td></tr>"
        "</table>"
        "<h2>Outlook</h2>"
    ),
}
verdicts = score(outputs, tests)

>>> verdicts
[Verdict(id='t1', result='pass', why=''),
 Verdict(id='t2', result='pass', why=''),
 Verdict(id='t3', result='fail', why="no cell matching '4.2' in any table")]

>>> s = summarize(verdicts, tests)["html"]
>>> s["score"], s["per_test_type"], s["per_tag"]["table"]
(0.6666666666666666, {'present': 1.0, 'order': 1.0, 'table_cell': 0.0}, {'score': 0.0, 'tests': 1, 'pages': 1})

A verdict is pass, fail, unsupported (the provider can't produce that output) or error (grading raised: a bug in the scorer, not the provider's fault), with why saying what went wrong. unsupported and error verdicts are left out of every score.

From the command line, with the outputs in a JSON file:

opb score --pred outputs.json --gold gold/<page_id>.json

It prints the verdicts. A score is a number over many pages: summarize in your code, or summary.json from opb benchmark. To grade your own model on the whole dataset, see The dataset.

Predict

With [harness] installed, predict with our provider adapters:

from omni_parse_bench import harness

record = harness.predict("datalab", "pages/<page_id>.pdf", {"html", "blocks"}, mode="balanced")
record["outputs"]       # {"html": "...", "blocks": [...]}
record["errors"]        # {output: why} for anything that failed
record["calls"]         # what was sent, cost, wall time

score(record["outputs"], gold.read("gold/<page_id>.json"))

From the command line:

opb predict --provider openai/gpt-5.6-sol --page pages/<page_id>.pdf --wants html
opb providers        # every provider, and its options

Every provider gets the same timeout per call and its most accurate settings by default. An output a vendor can't produce is recorded as unsupported, and its tests are left out of that vendor's scores rather than failed.

Documentation

• docs/API.md: every function and command, the dataset, and what a run writes;
• docs/design.md: each test type, tag and derivation, and the data layout;
• docs/building-tests.md: how the tests were built, and where each one's answer comes from;
• docs/providers.md: the provider adapters; and
• docs/tagging.md: how the tests' tags were made.

Found a test that's wrong, or a page you hold rights in? Open an issue.

License

The code is Apache 2.0: see LICENSE, and NOTICE for the third-party code it includes (KaTeX, MIT; a rewrite of olmOCR-bench's table parser, Apache 2.0).

The data's licenses are per source. Each source of the pages, its license and the credit it requires are in licenses/, one file per source. The tests are CC BY 4.0, except those converted from olmOCR-bench (ODC-BY 1.0) and fr-bench-pdf2md (MIT), and those quoting Wikipedia on three handwriting pages (CC BY-SA 4.0).

Citation

@misc{omniparsebench2026,
  title        = {Omni Parse Bench: A Document-Parsing Benchmark},
  author       = {{Datalab}},
  year         = {2026},
  howpublished = {\url{https://github.com/datalab-to/omni_parse_bench}},
}

Please also cite the upstream datasets whose pages or tests this benchmark uses; they are listed in licenses/.