Benchmark Almanac

Reducto

We track 3 benchmarks that Reducto made or helped make.

0 in use 2 with public code 2 with a leaderboard here

Benchmarks

Name Kind Released Best score Status
RD-TableBench Models extract 1,000 hand-labeled complex tables from document images, including scans, handwriting and merged cells. Scored by Levenshtein-based row and cell alignment. Documents, charts and text in images Nov 2024 57.4% System cards only
TreasuryQA Bench Agents search scanned Treasury volumes and read tables to find or calculate a number. A script checks whether the answer falls within the accepted range. Documents and spreadsheets Oct 2026 82.4% Not in cards
Table Judge Benchmark Vision language models judge whether rendered table images match candidate HTML. Controlled corruptions test false positives, true positives and injected-error detection. Documents, charts and text in images Nov 2025 94.9% Not in cards