Whether a model refuses or complies with medically harmful requests.

MedSafetyBench is a benchmark in healthcare AI. Whether a model refuses or complies with medically harmful requests. Scores are reported as: Harmfulness score across harmful request categories.

Audited by MedCheckHan et al., NAACL Industry Track 2024

Key facts

What it measuresWhether a model refuses or complies with medically harmful requests.
Who should careAnyone deploying a patient-facing tool.
Use it whenSafety review before launch.
Do not use it forMeasuring clinical accuracy.
Score scaleHarmfulness score across harmful request categories.
Grader methodunknown
First released2024
LanguagesEnglish
Use casessafety
CitationHan et al., NAACL Industry Track 2024
Statusactive

Frequently asked questions

What does MedSafetyBench measure?

Whether a model refuses or complies with medically harmful requests.

Who should use MedSafetyBench?

Anyone deploying a patient-facing tool. Safety review before launch.

What should MedSafetyBench not be used for?

Measuring clinical accuracy.

How are MedSafetyBench scores reported?

Harmfulness score across harmful request categories. Scores from different benchmarks are not comparable to each other.

Is MedSafetyBench independent?

No conflicts of interest have been verified for MedSafetyBench in this registry as of 2026-08-05. Absence of a recorded conflict means none has been verified, not that none exists.

Compare with

HealthBench

Physician-written rubrics score open-ended answers to realistic health conversations, including patient-facing

MedCalc-Bench

Whether a model computes clinical scores and formulas correctly from a patient note.

MEDEC

Whether a model can find and fix clinical errors already present in a note.

MEDIC

A framework covering several clinical application dimensions at once, including safety and reasoning, rather t

MediQ

Whether a model asks the right follow-up question when it does not have enough information, instead of guessin

MedR-Bench

Quantifies reasoning quality on real-world clinical cases, scoring the path rather than only the answer.

Last verified 2026-08-05. Governance and conflict data compiled by Healthcare Discovery from primary sources, each linked above. Seed inventory from Ma et al., Beyond the Leaderboard, ACL 2026. How we verify · All 63 entities