Preprints landing in the coming weeks

The first independent safety evaluation of frontier clinical AI.

Two coordinated preprints, one documenting the PsiBench methodology and three-tier framework, the other documenting what we found across 40 frontier models and 59,040 evaluations, are landing in the coming weeks.

The largest independent evaluation of clinical AI safety to date.

We evaluated 40 frontier language models from 10 providers against expert-authored medication-safety scenarios drawn from the standard 2,000+ U.S. hospitals are already audited against. The methods and findings are documented in two forthcoming preprints.

492
Validated safety scenarios
11
Safety categories
40
Frontier models
10
Model providers
59,040
Independent evaluations

No frontier model is ready to make
medication-safety decisions on its own.

The full results, including per-model performance, the three-tier evaluation framework, and the deployment-readiness analysis, are reserved for the preprints. The headline numbers are below.

9 / 40

Pass a basic deployment-readiness check

Defined on three operational criteria. Thirty-one of forty fall short on at least one.

19.5%

Of correct alerts cite the wrong reason

A model can detect a hazard and still attribute it to the wrong clinical category, roughly one in five times.

50+ pts

Spread on highest-stakes discrimination

On a single F1 number the field clusters within 14 points. At the tier level it splits apart.

6 โ€“ 82%

Specificity range, same headline F1

Several models reach 100% sensitivity at under 25% specificity, statistically the same as alerting on everything.

Two coordinated papers, landing in the coming weeks.

The methods paper introduces the PsiBench benchmark and the three-tier evaluation framework. The clinical paper applies it to 40 frontier language models and reports the field-wide findings.

Methods paper

A three-tier operational benchmark for evaluating large language models on hospital medication safety.

Introduces PsiBench (492 expert-authored scenarios across 11 safety categories) and a non-overlapping three-tier framework separating highest-stakes discrimination, the operational CDS regime, and category-correct alerting. Generalizes to any binary medication-safety classifier.

Preprint forthcoming. Request early access to read in full.
Clinical paper

No frontier LLM is ready for autonomous medication-safety decisions.

Applies the PsiBench benchmark and three-tier framework to 40 frontier language models from 10 providers (59,040 evaluations). Reports per-tier leaders, the field-wide attribution failure rate, and a deployment-readiness check that only 9 of 40 models pass.

Preprint forthcoming. Request early access to read in full.
Request early access โ†’