Dataset · TsinghuaC3I
MedXpertQA expert medical questions
TsinghuaC3I/MedXpertQA
4,460 challenging medical questions spanning text and multimodal tasks with clinical images.
What's inside
MedXpertQA is a challenging benchmark for evaluating expert medical knowledge and reasoning. It includes text-based and multimodal question-answering; the multimodal part uses structured clinical information and images. Totaling 4,460 questions across various specialties, body systems, and task types; language is English.
What you can do with it
- Evaluating medical LLMs
- Testing multimodal models on clinical images
- Benchmarking medical reasoning
- Tasks
- Table question answering, Question answering, Text generation
- Languages
- English
- Size
- 1K–10K rows
- Downloads this month
- 11.8K
- Likes
- 68
Models trained on it
How to load it
from datasets import load_dataset
ds = load_dataset("TsinghuaC3I/MedXpertQA")