MarathonDocBench
A benchmark for document parsing at extreme lengths, targeting scenarios of more than one hundred pages. It evaluates a model's ability to preserve information and structural consistency over very long contexts.
Timeline (last 90 days)
MarathonDocBench 是 Unlimited OCR 2 团队新提出的百页级文档解析基准，用于填补现有评测在超长文档区间上的空白。
Unlimited OCR 2 scored 77.1 on MarathonDocBench.
MarathonDocBench is an OCR evaluation benchmark oriented toward long documents.
A single benchmark score (such as 77.1) is insufficient to fully assess the real quality of long-document OCR; factors such as layout complexity and scan quality also need to be considered.
Baidu says Unlimited OCR 2 achieves SOTA performance on the tens-of-pages-level benchmark MPDocBench and the hundreds-of-pages-level benchmark MarathonDocBench.
MarathonDocBench targets extreme-length scenarios of over a hundred pages, examining the model's ability to retain information and maintain structural consistency over ultra-long contexts.
All Facts (6)
MarathonDocBench 是 Unlimited OCR 2 团队新提出的百页级文档解析基准，用于填补现有评测在超长文档区间上的空白。
50%UnverifiedUnlimited OCR 2 scored 77.1 on MarathonDocBench.
50%UnverifiedMarathonDocBench is an OCR evaluation benchmark oriented toward long documents.
50%UnverifiedA single benchmark score (such as 77.1) is insufficient to fully assess the real quality of long-document OCR; factors such as layout complexity and scan quality also need to be considered.
50%UnverifiedBaidu says Unlimited OCR 2 achieves SOTA performance on the tens-of-pages-level benchmark MPDocBench and the hundreds-of-pages-level benchmark MarathonDocBench.
50%UnverifiedMarathonDocBench targets extreme-length scenarios of over a hundred pages, examining the model's ability to retain information and maintain structural consistency over ultra-long contexts.
50%