ISL FinePDFs OCR

Pairwise VLM-as-judge OCR benchmark ranking OCR models on ISL FinePDFs document images with Bradley-Terry ELO scores.

3rows
eloprimary metric
2026-05-06sampled

Metadata

Metrics

ELO, Win%, Wins, Losses (lower is better), Ties, ELO Low, ELO High

Latest Results

Rows are parsed from the public Hugging Face dataset-server rows API. ELO rankings come from OCR Bench pairwise VLM-as-judge comparisons.

Rank Subject ELO Model Match Provenance Sampled
1 rednote-hilab/dots.ocr 1613 Imported 2026-05-06
2 deepseek-ai/DeepSeek-OCR 1462 Imported 2026-05-06
3 zai-org/GLM-OCR 1425 Imported 2026-05-06