VCT (SecureBio)
Science
SecureBio's own runs of the Virology Capabilities Test, 322 multiple-response virology troubleshooting questions, 221 with images. These are the same questions as vct, whose rows come from xAI model cards, under a different protocol. Evaluated in Inspect with zero-shot prompting, usually over 10 runs. SecureBio doesn't state how refusals enter accuracy. Claude Opus 5.5 prints 61.4% with 79.8% refused, more than it could score if refusals counted as wrong, so accuracy appears to cover answered questions only and each model's denominator differs. These scores aren't comparable with other runs of the test, which is why this is a diagnostic.
Benchmark source: Securebio