Insights

Method documentation, result analysis, and guidance on interpreting results for your own caseload.

The archive

Showing 1–5 of 5 posts

25 Sept 2026Methodology

A 95% Accurate AI Can Still Kill Your Patient

Why we stopped trusting percentages when we built a benchmark for veterinary AI

MGMichael Gerges, DVM· 5 min read

22 Sept 2026Methodology

A refusal is not a safe answer

When a model declines a safety question we score it as though it never saw the item. Why a refusal earns no credit, and what that changes about the ranking.

7 min read

16 Sept 2026Methodology

The held-out set that wasn't

A fixed selection seed quietly turned our held-out set into a fixed exam every vendor could prepare for. How the bug works, and why a bigger paper is not the fix.

6 min read

10 Sept 2026Guides

How to read an AI benchmark before you trust it

The questions to ask of any leaderboard before you let it decide a purchase, including the ones we would rather you asked us.

8 min read

05 Sept 2026Field

The benchmark veterinary medicine was asked for, and never got

Two veterinary colleges called for an oversight body for veterinary AI in 2025. It still does not exist. What formed in the gap, and who pays for it.

9 min read

Scope. VetEval measures model performance on veterinary examination items. It is not clinical advice, and no result here licenses a model for clinical use.