The IDA (Intelligent Document Analysis) software suite claims to offer outstanding Optical Character Recognition (OCR) accuracy by leveraging patented core technology. To provide quantified evidence for this claim, we conducted a benchmark Analysis across leading commercial and open-source OCR solutions. In this update, we evaluate the LLM-based engines Gemini 2.0 Flash, GPT-4o, and Mistral OCR.

Open-source engines: easyOCR, MMOCR, PaddleOCR, Tesseract

Commercial engines: Amazon, Azure, Google, Planet AI

Multimodal LLM-based engines: GPT-4o (OpenAI), Mistral OCR, Gemini 2.0 Flash (Google)

By downloading the benchmark, you gain access to the 16-page white paper with the results.