We are building the world's most specialized open-source medical language models — freely available to every researcher, clinician, and developer on the planet.
OpenMedLLM is an open-source initiative by DeepCog.ai, a specialized AI research company headquartered in Chennai, India, with a research lab operating in proximity to IIT Madras and Anna University — two of the world's leading centers for AI and computational biology research.
We build domain-specific large language models trained on medical databases, clinical variant repositories, biomedical literature, and expert-annotated preference data. Every model we build is released openly under the Apache 2.0 license, with full weights, training code, and datasets available to the global research community.
Every model, dataset, and training script we create is released openly. Medical AI advances when everyone can build on each other's work.
We never release a model without rigorous clinical validation. In medicals, a wrong variant classification isn't a UX issue — it's a patient safety issue.
Our lightest models run on modest hardware. Medical AI should serve researchers in Mumbai, Nairobi, and São Paulo — not just Stanford and UCSF.
AI augments the clinical geneticist — it never replaces them. Every tool we build keeps the human expert in control of the final interpretation.
Download models, contribute datasets, and help build the future of medical medicine.