CUREMED-BENCH

CUREMED-BENCH is a multilingual medical reasoning benchmark dataset, designed for evaluating and fine-tuning models on medical tasks across diverse languages.

Data

set Summary

• Languages: Spans 13 languages, including Amharic, Bengali, French, Hausa, Hindi, Japanese, Korean, Spanish, Swahili, Thai, Turkish, Vietnamese, and Yoruba.
• Splits: Includes train, test, and validation splits, each with separate CSV files per language.
• Usage: Intended for research in multilingual medical AI.

Structure

• train_dataset/: Training data CSVs.
• test_dataset/: Test data CSVs.
• val_dataset/: Validation data CSVs.

Sources

• Repository: https://github.com/AikyamLab/cure-med
• Paper: https://arxiv.org/abs/2601.13262
• Demo: https://cure-med.github.io/

Citation

BibTeX:

@article{onyame2026cure,
  title={CURE-Med: Curriculum-Informed Reinforcement Learning for Multilingual Medical Reasoning},
  author={Onyame, Eric and Ghosh, Akash and Baidya, Subhadip and Saha, Sriparna and Chen, Xiuying and Agarwal, Chirag},
  journal={arXiv preprint arXiv:2601.13262},
  year={2026}
}