Skip to content

Back to the catalog

XBOW: GPT-5.5 and Opus 4.7 in Offensive Security Workflows
xbow.comwriteup

XBOW: GPT-5.5 and Opus 4.7 in Offensive Security Workflows

XBOW shares internal benchmark results comparing GPT-5.5, GPT-5, GPT-5.4, and Opus 4.6/4.7 on autonomous vulnerability-finding agents, using frozen vulnerable versions of open source apps. GPT-5.5 cut the miss rate from GPT-5's 40% to 10%, dramatically improved both black-box and white-box exploitation, logged into target apps in half the iterations of competitors, and gave up on dead-end paths twice as often as prior models — a key metric for agentic offensive tooling.

TL;DR XBOW's benchmarks show GPT-5.5 slashing vulnerability miss rates to 10% and improving autonomous pentest agent efficiency over GPT-5/Opus 4.6.
Open original ↗

Technologies named here

Techniques named here

  • Black box penetration testing
  • White box penetration testing
  • Automated vulnerability discovery and exploitation
  • Credential/login testing
  • Persist-or-pivot decision making

Related in Offensive Security