What it measures

4,460 specialist-board-level medical questions across 17 specialties and 11 body systems (including radiology/pathology images).

Reading the measurement

Accuracy is only part of the picture. Inspect unsupported answers, abstentions, language coverage and the evidence available to the model. A confident response can still be wrong, and a broad average can hide a specialist gap.

Published evaluation

Reasoning and understanding questions use the source’s published aggregate. A dagger marks results evaluated on a sampled subset; those source model labels are retained.

Where this fits in an evaluation suite

Read accuracy together with error and abstention rates. Domain averages can hide language gaps, specialist failures, and poorly calibrated confidence.

Start with a capability hypothesis, then choose a second test with a different failure mode. Keep the checkpoint, reasoning effort, agent tools, sampling policy, and scoring procedure fixed when measuring a training intervention. Record failures and missing outputs in the denominator.

A reproducible comparison

Reasoning and understanding questions use the source’s published aggregate. A dagger marks results evaluated on a sampled subset; those source model labels are retained.

Publish the exact run configuration, item coverage, score units, and available uncertainty. If a source supplies only an aggregate or a single run, preserve that limit. A new model name or a rounded score alone does not establish a capability improvement.

PRIMARY EVIDENCE & FURTHER READING

  1. MedXpertQA: Benchmarking Expert-Level Medical Reasoning and Understanding — published result table ↗

    Yuxin Zuo, Shang Qu, Yifei Li, Zhangren Chen, Xuekai Zhu, Ermo Hua et al. · 2025

    All 9 authors

    Yuxin Zuo, Shang Qu, Yifei Li, Zhangren Chen, Xuekai Zhu, Ermo Hua, Kaiyan Zhang, Ning Ding, Bowen Zhou

Reviewed 2026-10-26. Explanation and interpretation by RLTasks. Benchmark ownership remains with its original authors.