EPFLiGHT/Meditron3-Qwen2.5-7B

For a convenient overview and download list, visit our model page for this model.

A domain-optimized reasoning model built on DeepSeek-R1-Distill-Qwen-32B, refined through a multi-stage pipeline of GPTQ quantization-aware training and QLoRA fine-tuning. Achieves 84% on MedQA — within 4 points of GPT-4o — in a ~20GB package that fits on a single L40/L40s GPU.

> NEXUS domain specialist for medical Q&A and clinical reasoning — lightweight & uncensored.

Model Size: 7B