This site is a work in progress and has not been widely shared. Content may contain errors. Feedback is welcome.
This site is undergoing review. Some annotations were human-generated, some AI-generated — all are being verified.
Back to datasets

FinQA

FinQA: A Dataset of Numerical Reasoning over Financial Data

AI-focusedPublicNeither
Visit Dataset
Specific Type
AI benchmarking
Dataset Type
Cross-sectional
Institution
University of California Santa Barbara; Salesforce Research; Carnegie Mellon University
Institution Type
Academia; Industry
Level of Focus
Task capability
Most Granular Level
Individual financial question level
Perspective
Neither
Time Coverage
2021-present
Frequency
Static benchmark
Sample Size
8281 financial QA pairs
Geographic Detail
National (US)
Occupational Classification
Not specified
Industrial Classification
S&P 500 companies
Other Classification
Financial reasoning tasks
Key Variables
Financial reasoning accuracy; numerical computation; multi-step reasoning over financial documents
AI/Tech Tracking
Complex numerical reasoning over financial reports and earnings data
Access Details
Available on GitHub and Papers with Code
Notes
Expert-annotated by 11 finance professionals; requires multi-step reasoning over tables and text

Key Papers

Chen et al. (2021) EMNLP