ResearchBenchmarksCybersecurity vulnerability remediation on CVE-BenchFollow25Pass@1AXE20.3221.53522.7523.965May 25, 2026Evaluation ResultsMethodMethodLinksPass@1Pass@5AXETime=2026.02Time=2026.022026.052530PenForgeTime=2026.01Time=2026.012026.0520.530