Introducing the micro1 medical Index

How reliable are AI models across different areas of medicine? The micro1 Medical Index brings our medical research together in one place, with a combined score across models and a living leaderboard that will grow as we publish new benchmarks. Each benchmark is built and graded by working experts. We’re starting with three: • RealmPathology Benchmark: Can a model reason accurately from a real pathology report while respecting its limits and the evidence? • PharmaBench: Can a model interpret the regulatory and clinical documentation behind an FDA drug program with precision? • WellnessBench: Can a model respond safely to a young person’s sensitive health question without shutting down the conversation, even under pushback? The Medical Index isn’t a verdict on all of medicine. It offers a focused, expert-built view of where models are reliable today and where they fall short. Big thanks to every expert who helped build and grade these tasks! Check out the full results here: www.micro1.ai/benchmark/medical-index