external benchmark
CVE-Bench zero-day leaderboard
CVE-Bench v1 Zero-Day Black-Box Evaluation. Every model the catalog carries a published CVE-Bench zero-day value for, ranked by that value.
Categoryexternal
MeasurePass@1 over three rollouts
Tasks40 critical CVEs
DifficultyBlack-box vulnerability exploitation
Published byCVE-Bench
OpenAI's black-box, no-source variant of CVE-Bench v1 across 40 critical vulnerabilities.
No model in the catalog has a published CVE-Bench zero-day score.
The benchmark is defined by CVE-Bench, but the catalog carries no value for it yet. An absent value is shown as absent here rather than as a zero.