We present the HIPE-OCRepair-2026, an ICDAR competition focused on LLM-assisted OCR post-correction for historical documents. OCR post-correction remains a long-standing challenge in digital heritage, with large-scale collections of digitized documents suffering from legacy OCR errors, while large-scale re-digitization remains impractical. Large language models (LLMs) offer a significant opportunity to revisit this challenge, yet their effectiveness across languages, document types, and noise conditions, along with their tendency to hallucinate, is not yet fully understood.
The HIPE-OCRepair-2026 has two main objectives: (i) to evaluate the capabilities of modern OCR post-correction systems, and (ii) to provide a reproducible evaluation framework based on the HIPE-OCRepair-2026 dataset, a harmonized multilingual resource consolidating existing and newly curated historical datasets. Participants were tasked with correcting noisy OCR transcripts from historical newspapers and printed works in English, French, and German (17th-20th century), working at the level of coherent transcription units (paragraphs or articles) without access to source images.
The evaluation adopted a retrieval-oriented rather than diplomatic scoring approach, reflecting the practical use case of search and access over digitized collections. Four teams submitted systems ranging from zero-shot prompting to continued pre-training and fine-tuning, providing insights into the merits of different adaptation strategies. Results indicate that modern LLM-assisted systems can significantly improve OCR quality, but performance varies across datasets, languages, and noise levels. Over-correction on low-noise inputs emerges as a recurring challenge, highlighting the importance of evaluation beyond character error reduction. The dataset, scorer, and evaluation pipeline are publicly released to support future research.
Blogger's Review: HIPE-OCRepair-2026 showcases the potential of LLMs in OCR post-correction for historical documents, despite various challenges. The innovative and diverse evaluation criteria lay a foundation for future related research, with a particular focus on balancing correction quality and over-correction issues. Such competitions drive advancements in OCR technology, warranting continued attention from the tech community.