Composite score from the Nanonets IDP Leaderboard — unweighted mean of three sub-benchmarks: OlmOCR (OCR quality, 7,010 unit tests), OmniDocBench (document parsing, 18K+ samples), and IDP Core (extraction + VQA, 6,406 samples). Covers 24 models including specialist tools and frontier LLMs.
Most comprehensive enterprise document processing benchmark; aggregates 3 complementary evaluations; includes specialist tools alongside LLMs; actively maintained.
Unweighted mean of 3 sub-benchmarks may not match specific enterprise priorities; some specialist tools are purpose-built for specific tasks.
Higher is better. Unweighted mean of OlmOCR, OmniDoc, and IDP Core scores, 0-100.
Max score: 100
Unweighted mean of 3 sub-benchmarks may not match specific enterprise priorities.
| # | Model | Score |
|---|---|---|
| 1 | 85.9 | |
| 2 | 83.5 | |
| 3 | 83.5 | |
| 4 | 82.8 | |
| 5 | 82.0 | |
| 6 | 82.0 | |
| 7 | 81.8 | |
| 8 | 81.6 | |
| 9 | 81.5 | |
| 10 | 80.7 | |
| 11 | 80.4 | |
| 12 | 79.6 | |
| 13 | 76.7 | |
| 14 | 75.2 | |
| 15 | 72.5 | |
| 16 | 71.5 | |
| 17 | 71.2 | |
| 18 | 71.2 | |
| 19 | 69.5 | |
| 20 | 68.0 | |
| 21 | 64.2 | |
| 22 | 62.6 | |
| 23 | 57.8 | |
| 24 | 54.8 | |
| 25 | 53.9 | |
| 26 | 46.5 | |
| 27 | 41.9 |