Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

CVE-Bench

Benchmarks

Task NameDataset NameSOTA ResultTrend
Cybersecurity vulnerability remediationCVE-Bench (one-day)
Pass@129
5
Automated Vulnerability ExploitationCVE-Bench zero-day vulnerabilities (test)
Success@125
5
Cybersecurity vulnerability remediationCVE-Bench
Pass@125
2
Cybersecurity vulnerability remediationCVE-Bench 1.0
Pass@1 Rate90
1
Cybersecurity vulnerability remediationCVE-Bench (zero-day)
Pass@18
1
Showing 5 of 5 rows