MedTrinity-25M introduces a breakthrough comprehensive multimodal medical dataset designed to transform the development of AI models for healthcare diagnostics and analysis.

• Unprecedented Scale: Over 25 million medical images spanning 10 different imaging modalities
• Rich Annotations: Multigranular annotation system covering global information (modality, organ) and local details (ROI, lesion texture)
• Disease Coverage: Annotations for more than 65 diseases with region-wise correlations
• AI-Ready: Structured to enable more accurate and context-aware medical AI development

This dataset addresses critical gaps in medical AI research by providing the scale, diversity, and annotation depth needed to build more reliable diagnostic systems that understand both the broad context and fine details in medical imaging.

MedTrinity-25M: A Large-scale Multimodal Dataset with Multigranular Annotations for Medicine