Four-Factor Integrity Breakdown & Primary-Source Evidence
Access Model: Strictly Private
"This dataset remains private at all times, and is the only dataset used for the proprietary benchmark results we publish. It is private to prevent leakage into the training data of foundation models."
"A patch submission is marked as a pass if the edited source tree compiles cleanly, the original PoC no longer triggers the sanitizer error, and hidden holdout checks match the maintainer-fixed reference."
Originality
Weight 25%
0/10
Reuse Profile: High Reuse (>75%)
"Agents are tested on OSS-Fuzz regressions from ARVO images. In cybersecurity, fuzzing is an automated testing technique that feeds many inputs into a program to trigger crashes and memory-safety bugs."
Freshness
Weight 20%
10/10
Initial Release: 2026-09Latest Refresh: 2026-11Rolling Updates: No
"CyberBench v1.1 supersedes v1 because the results were rerun with offline task sandboxes and a hardened clean-room grader. It keeps the 60-task PoC track and publishes a 56-task Patch roster containing the tasks that can be graded reliably under that setup."
Integrity: Independent Offline (Leak-Verified)Sandbox: offline isolated
"CyberBench v1.1 supersedes v1 because the results were rerun with offline task sandboxes and a hardened clean-room grader. It keeps the 60-task PoC track and publishes a 56-task Patch roster containing the tasks that can be graded reliably under that setup."
Frontier Model Standings on CyberBench v1.1
Top 15 active models tracked
Benchmark maintainers can embed this live SVG badge in GitHub READMEs or documentation:
[](https://benchaudit.org/benchmarks/cyberbench/)