OCR evidence benchmark
Cross-benchmark scores, historical cohorts and documented prices.
| Model | Rank index | DPB | DrDoc | IDP | MDP | MORE | NDI | olm | OmniDoc | Pure | Real5 | Wild | $/1k pages |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| TeleOCR | 99.4 | — | 68.0† | — | — | — | — | — | 96.9 | 78.5 | 90.7 | 88.5† | — |
| OvisOCR2 | 90.7 | — | 59.3† | — | — | — | — | — | 96.5 | 75.1 | 92.3 | 87.9† | — |
| Chandra 2 API | 74.0 | 0.791* | — | — | 82.4* | — | 0.773 | 86.7 | — | — | — | — | — |
| Gemini 3 Pro | 73.3 | — | — | 81.8 | — | — | — | 77.7 | 92.9 | — | 89.2 | — | $19.0 |
| Chandra 2 open | 73.2 | 0.791 | — | — | 82.4 | — | — | 85.9 | — | — | — | — | — |
| PaddleOCR-VL-1.6 | 69.8 | — | 55.1† | — | 83.0 | — | — | — | 96.3 | 59.3 | 93.2 | 87.4† | — |
| MinerU2.5-Pro | 68.2 | — | 62.3† | — | 74.3 | — | — | — | 95.8 | 70.1 | 88.9 | 87.3† | — |
| Qwen3-VL-235B | 59.1 | — | — | 80.0 | — | — | — | 76.8 | 89.8 | — | 88.9 | — | — |
| GPT-5.2 | 55.8 | — | — | 77.4 | — | — | — | 79.1 | 86.6 | — | 78.7 | — | $21.4 |
| dots.ocr | 53.4 | — | — | — | 79.7 | 84.3 | — | 79.1 | 90.8 | 64.5 | 86.4 | 81.8† | — |
| MinerU2.5 | 49.5 | 0.816 | — | — | — | 48.9 | — | — | 93.0 | 67.7 | 85.6 | — | — |
| HunyuanOCR | 41.9 | 0.513 | — | — | 68.6 | 92.4 | — | — | 90.0 | 60.6 | — | — | — |
| MonkeyOCR-pro-3B | 38.0 | 0.826 | — | — | 53.6 | — | — | — | 88.6 | 55.4 | 79.5 | — | — |
| GLM-OCR | 37.4 | 0.735 | — | 54.9 | 68.8 | — | — | 68.4 | 95.2 | 63.3 | 90.3 | 85.1† | — |
| DeepSeek-OCR 2 | 29.1 | 0.740 | — | — | — | — | — | — | 90.3 | 49.5 | 73.0 | — | — |
| DeepSeek-OCR | 23.0 | — | — | — | 54.5 | 82.9 | — | 75.7 | — | 47.0 | 74.0 | — | — |
* open-model floor · † qualified / author-reported · — no comparable result
| Model | USD / 1,000 pages | Basis | As of |
|---|---|---|---|
| Mistral OCR 4.1 | $4.0 | Official page price | 2026-09-30 |
| Datalab Convert / Chandra 2.1 | $4.0 | Official page price | 2026-09-30 |
| Extend Parse | $6.3 | Official page price | 2026-09-30 |
| Reducto r-1 Parse | $10.0 | Official page price | 2026-09-30 |
| Nanonets OCR-3 | $10.0 | IDP listing | 2026-09-30 |
| GPT-6 Astra (low) | $70.6 | NDI harness | 2026-09-30 |
| GPT-5.6 Terra | $15.5 | NDI harness | 2026-09-30 |
| GPT-5.2 | $21.4 | IDP Leaderboard | 2026-09-30 |
| GPT-5-Mini | $3.0 | IDP Leaderboard | 2026-09-30 |
| GPT-5-Nano | $0.6 | IDP Leaderboard | 2026-09-30 |
| GPT-4.1 | $13.6 | IDP Leaderboard | 2026-09-30 |
| Anthropic Opus 4.7 | $71.4 | NDI harness | 2026-09-30 |
| Claude Opus 4.6 | $40.8 | IDP Leaderboard | 2026-09-30 |
| Claude-Sonnet-4.6 | $24.4 | IDP Leaderboard | 2026-09-30 |
| Claude Haiku 4.5 | $8.2 | IDP Leaderboard | 2026-09-30 |
| Gemini 3 Pro | $19.0 | IDP Leaderboard | 2026-09-30 |
| NDI OCR (low) | $2.7 | NDI harness | 2026-09-30 |
| NDI OCR (high) | $6.5 | NDI harness | 2026-09-30 |
| LlamaParse Cost Effective | $3.8 | NDI harness | 2026-09-30 |
| Databricks Parse (high) | $6.1 | NDI harness | 2026-09-30 |
| Azure Document Intelligence | $10.0 | NDI harness | 2026-09-30 |
| LlamaParse Agentic | $13.8 | NDI harness | 2026-09-30 |
| Mistral OCR 4.1 | $4.0 | vendor_list | 2026-10-13 |
| Datalab Convert Accurate | $10.0 | third_party_listing | 2026-10-05 |
Methodology, provenance & sources
Each mature independent benchmark contributes one equal vote to the mean percentile-rank index. A benchmark is included only when the historical cohort has six scored models from three organizations. At least three core families are required for a Rank index; 4+ primary families is the default display filter.
Author-reported, internal, and small-sample results are shown but are excluded from the core Rank unless corroborated by comparable evaluation. Dates are model release dates, not benchmark publication dates. Scores and prices are source-specific; historical views are retrospective cohorts, not archived leaderboards.
Some older records were migrated from an existing research snapshot; their source links are preserved, but they have not all been independently re-run. Missing model releases are omitted from historic cohorts until the current snapshot.
Official OmniDocBench comprehensive table — Current displayed table labelled v1.6_full; repo update log mentions v1.7.
PureDocBench leaderboard — Source-traceable clean / digital degradation / real degradation.
Official MDPBench leaderboard — Public and held-out private results; 17 languages.
Nace AI seven-benchmark aggregate — Seven-benchmark same-harness normalized aggregate; vendor operated, may overlap underlying tests.
TeleOCR public project comparison — TeleOCR team reports its comparison with competitors using Wild-OmniDoc and a limited four-model Dr.Doc run; not an independent reproduction.
olmOCR / Datalab / Nanonets published comparisons — Provenance is specified per entry; different runs and harnesses.
DocParsingBench — Community document-parsing aggregate.
Nanonets IDP benchmarking — Independent model comparisons in vendor-operated harness.
Real5-OmniDocBench — Real-world degradations of OmniDoc v1.5 pages; correlated with OmniDoc.
MORE 149-language benchmark — Multilingual real document parsing.
PolyOCR-Venus technical report — MDP public; no private score; weights unavailable per project release.
Xiaomi-OCR-0 technical report — OCR authors evaluated their own released weights; external layout detector for parsing.
OpenOCR 20-page hard set — Published 2026-10-05; 20 pages and 141 rule tests, uncertainty intervals overlap.
Datalab Chandra 2.1 release — 90-language and complex table scores; not comparable with public suites.
Mistral OCR pricing — OCR 4.1 is 4 per 1000 pages, 0.40 batch; annotated 5.
OCR research snapshot 2026-09-30 — Prior research import; not a new independent rerun. Underlying source links preserved.
IndicOCR v1.6 English subset — English subset is not v1.6_full and must not silently replace it.