ParseBench — Apache License 2.0

Source: https://huggingface.co/datasets/llamaindex/ParseBench
Copyright: LlamaIndex.
Suites: parsebench_docs, its rotated copies in rotation_skew, and the synth_copies pages rebuilt
from it.

Pages are cut from upstream's documents; their annotations are not used, and this benchmark's tests
were made for it. Upstream describes the documents as sourced from public online channels.

Please cite: Zhang B. et al. (2026). ParseBench: A Document Parsing Benchmark for AI Agents.
arXiv:2604.08538.

Full license text: https://www.apache.org/licenses/LICENSE-2.0
