Return to analysisSOURCE NOTE / BENCHMARK / UC BERKELEY AND COLLABORATORS PROVENANCE DETAILS
CyberGym-E2E
A benchmark spanning discovery, reproduction and patching, with executable checks across the vulnerability lifecycle.
Open original material01 / THE PUBLICATIONUC Berkeley and collaborators
Paper and reproducible dataset
02 / ITS RELATIONSHIP HEREResearch benchmark
Record reviewed Sep 25, 2026
03 / THE CLAIM BOUNDARYPassing a repair check is bounded by that check. Benchmark outcomes alone do not establish every security or functional property.
- Published or updated
- Jun 3, 2026
- Record reviewed
- Sep 25, 2026
- Available material
- Paper and reproducible dataset