Benchmark RadarFinBalance
FinBalance is a multi-document accounting reconciliation benchmark built from 710 source-document bundles across eight industries, three period types, and five difficulty levels. It evaluates models on producing journal entries, balance sheets, and inconsistency labels, with a deterministic ledger for ground truth.
- Released
- 2026-06-14
- Readiness
- Runnable
- Primary field
- Finance & Economics
Why it matters
FinBalance addresses the gap in evaluating accounting reasoning from source documents rather than prepared financial statements. It provides a reproducible and auditable benchmark for document-grounded financial reasoning, with expert validation of the design.
Motivation
Existing financial-NLP benchmarks mostly evaluate prepared artifacts such as filings, tables, or extracted values.
Primary resources
Benchmark Radar records only publicly supported details and links back to primary sources for verification.