oi-OCR / README.md
Thiago-cs's picture
Update README.md
ad20e64 verified
---
tags:
- document-parsing
- ocr
- pdf
- parsebench
- enterprise-documents
license: apache-2.0
language:
- en
- ar
---
# oi-OCR
**oi-OCR** is Open Innovation AI's document-parsing tool. It extracts structured Markdown, layout, tables, and chart data from PDFs for downstream RAG ingestion, agentic workflows, and document understanding tasks.
## ParseBench Results (April 2026)
| Dimension | Score | Rank on the public leaderboard |
|---|---:|---|
| **Charts** | **78.48** | **#1 of 47** |
| Tables | 87.06 | #9 |
| Content Faithfulness | 87.24 | #18 |
| Semantic Formatting | 65.65 | #6 |
| Visual Grounding | 68.71 | #6 (tied with Reducto) |
| **Overall (mean of 5)** | **77.43** | **#2 of 47** |
Evaluated on the full [ParseBench-Full](https://huggingface.co/datasets/llamaindex/ParseBench) suite — 2,037 single-page PDFs across chart, layout, table, and text groups.
**oi-OCR is #1 on the Charts dimension** — ahead of LlamaParse Agentic (78.11), Reducto Agentic (73.40), Google Gemini 3 Flash Thinking High (64.79), Anthropic Opus 4.7 (55.84), and OpenAI GPT-5.5 Reasoning Medium (65.53).
On Overall, only LlamaParse Agentic ranks higher.
Structured eval data: [`.eval_results/parsebench.yaml`](./.eval_results/parsebench.yaml).
## Evaluation methodology
- **Benchmark**: [ParseBench-Full](https://huggingface.co/datasets/llamaindex/ParseBench) — 2,037 single-page PDFs from real enterprise documents (insurance, finance, government, scientific, etc.)
- **Evaluator**: official [`parse-bench`](https://github.com/run-llama/ParseBench) CLI
- **Scoring mode**: rule-only (`LLAMACLOUD_BENCH_LLM_NORMALIZATION=off`) — stricter than the leaderboard's default judge mode.
-
## Public leaderboard
Full benchmark comparison across all 47 entries: [parsebench.ai](https://www.parsebench.ai/)
## About
[Open Innovation AI](https://openinnovation.ai/) builds enterprise AI tools for the GCC and beyond, with first-class English and Arabic document support.