Queryome: Orchestrating Retrieval, Reasoning, and Synthesis across Biomedical Literature

Abstract The rapid expansion of biomedical literature has made comprehensive manual synthesis increasingly difficult to perform effectively, creating a pressing need for AI systems capable of reasoning across verified evidence rather than merely retrieving it. However, existing retrieval-augmented generation (RAG) methods often fall short when faced with complex biomedical questions that require iterative reasoning and multi-step synthesis. Here, we developed Queryome, a deep research system consisting of specialized large language model (LLM) agents that can adapt their orchestration dynamically to a wide range of queries. Using a hybrid semantic–lexical retrieval engine spanning 28.3 million PubMed abstracts, it performs iterative, evidence-grounded synthesis. On the MIRAGE benchmark, Queryome achieved 88.98 % accuracy, surpassing prior systems by up to 14 points, and improved reasoning accuracy on the biomedical Human’s Last Exam (HLE) subset from 15.8% to 19.3%. Moreover, in a task for constructing a review article, it earned the highest composite score in comparison with Deep Research from OpenAI, Google, Perplexity, and Scite.AI, reflecting its strong literature retrieval and synthesis capabilities. Competing Interest Statement DK is a founding member of Intellicule LLC. Data and Code Availability The source code of Queryome is available at https://github.com/kiharalab/queryome. In addition, an application that runs on MacOS (Intel and Apple Silicon) and Windows is available at: https://www.queryome.app/. Benchmark results can be downloaded from: https://kiharalab.org/queryome/benchmark_data.tar.gz.