Evaluating Medical LLMs by Levels of Autonomy: A Survey Moving from Benchmarks to Applications

TL;DR

This survey redefines the evaluation of medical large language models by levels of autonomy, emphasizing risk-aware, application-oriented assessment over traditional benchmark scores to better ensure safe clinical deployment.

Contribution

It introduces a levels-of-autonomy framework for evaluating medical LLMs, linking benchmarks to clinical actions and risks, and guides credible, risk-aware assessment for real-world use.

Findings

Aligns benchmarks with autonomy levels and associated risks

Proposes a level-conditioned blueprint for evaluation and reporting

Moves evaluation focus from scores to clinical safety and reliability

Abstract

Medical Large language models achieve strong scores on standard benchmarks; however, the transfer of those results to safe and reliable performance in clinical workflows remains a challenge. This survey reframes evaluation through a levels-of-autonomy lens (L0-L3), spanning informational tools, information transformation and aggregation, decision support, and supervised agents. We align existing benchmarks and metrics with the actions permitted at each level and their associated risks, making the evaluation targets explicit. This motivates a level-conditioned blueprint for selecting metrics, assembling evidence, and reporting claims, alongside directions that link evaluation to oversight. By centering autonomy, the survey moves the field beyond score-based claims toward credible, risk-aware evidence for real clinical use.