FinixDocBench: Financial Document Parsing Benchmark for Noisy and Digital Documents

FinixDocBench is a compliance-reviewed public subset of a financial-domain document parsing benchmark. It focuses on document parsing conditions common in real financial workflows but underrepresented in clean-document benchmarks, such as digitally native insurance clauses and noisy camera-captured medical bills. The dataset was created by inclusionAI and last updated on June 10, 2026.

Use Cases

• Benchmarking document parsing models based on noisy camera-captured documents.
• Training models to parse digitally native insurance clauses.
• Evaluating the robustness of financial information extraction systems on underrepresented document types.

Strengths

• Focuses on real-world financial document parsing conditions underrepresented in other benchmarks.
• The dataset is described as a compliance-reviewed public subset, suggesting some quality control.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and file formats are unknown, which may limit suitability assessment.

Provenance

Source

huggingface

Collection Method

Introduced in the technical report 'FinixDoc: Rethinking Financial Document Parsing Beyond Saturated Benchmarks'.

Freshness

Last updated 2026-06-10 11:54:51; freshness should be verified.