Skip to content

Dataset · SakanaAI

JFinBench Japanese financial benchmark

SakanaAI/JFinBench

An exam-based benchmark for evaluating LLMs in the Japanese financial domain.

What's inside

Includes 321 Japanese question-answering and visual question-answering examples drawn from 2026 exams across six financial domains. The dataset contains fewer than 1K entries and is strictly intended for model evaluation.

What you can do with it

  • Evaluating and benchmarking LLMs in the Japanese financial domain
  • Testing vision-language multimodal systems
Tasks
Question answering, Visual question answering
Languages
Japanese
Size
up to 1K rows
Downloads this month
745
Likes
3
Access
on request on Hugging Face

How to load it

from datasets import load_dataset

ds = load_dataset("SakanaAI/JFinBench")