olmOCR: Unlocking Trillions of Tokens in PDFs with Vision Language Models

Benchmark Model Rank Results
optical-character-recognition-ocr-on-olmocr-bencholmOCR#10Accuracy: 75.5
question-answering-on-natural-questionsOLMo-2-7B-1124 (olmOCR-peS2o)#46EM: 29.1
common-sense-reasoning-on-winograndeOLMo-2-7B-1124 (olmOCR-peS2o)#136Accuracy: 58.0