The paper introduces BiMediX, the first bilingual medical mixture of experts LLM designed for seamless interaction in both English and Arabic. BiMediX facilitates a wide range of medical interactions, including multi-turn chats, multiple-choice question answering, and open-ended question answering. The authors propose a semi-automated English-to-Arabic translation pipeline with human refinement to ensure high-quality translations. They also introduce a comprehensive evaluation benchmark for Arabic medical LLMs and an extensive Arabic-English bilingual instruction set named BiMed1.3M, covering 1.3 million diverse medical interactions. BiMediX outperforms state-of-the-art models such as Med42 and Meditron by average absolute gains of 2.5% and 4.1%, respectively, while operating 8 times faster. Additionally, it surpasses the generic Arabic-English bilingual LLM, Jais-30B, by average absolute gains of 10% on Arabic medical benchmarks and 15% on bilingual evaluations. The project page with source code and trained model is available at <https://github.com/mbzuai-oryx/BiMediX>.The paper introduces BiMediX, the first bilingual medical mixture of experts LLM designed for seamless interaction in both English and Arabic. BiMediX facilitates a wide range of medical interactions, including multi-turn chats, multiple-choice question answering, and open-ended question answering. The authors propose a semi-automated English-to-Arabic translation pipeline with human refinement to ensure high-quality translations. They also introduce a comprehensive evaluation benchmark for Arabic medical LLMs and an extensive Arabic-English bilingual instruction set named BiMed1.3M, covering 1.3 million diverse medical interactions. BiMediX outperforms state-of-the-art models such as Med42 and Meditron by average absolute gains of 2.5% and 4.1%, respectively, while operating 8 times faster. Additionally, it surpasses the generic Arabic-English bilingual LLM, Jais-30B, by average absolute gains of 10% on Arabic medical benchmarks and 15% on bilingual evaluations. The project page with source code and trained model is available at <https://github.com/mbzuai-oryx/BiMediX>.