OCR model suggestion

With the epic growth of gen AI models, I found that there are yet to have OCR models that could reach high accuracy rate like Google OCR.

Any open source OCR model to recommend? Apart from pytesseract - which I think has more room to grow.