0

SEC-Bench Pro

Agentic cybersecurity benchmark covering reverse engineering and security analysis tasks.

Domain
cybersecurity
Published
Sep 2026
Notable for
Cybersecurity benchmark used at the GPT-6 Astra launch.

Cite

Notes

Only stored in your browser.

Attribution

Leaderboard scores
OpenAI
Attribution policy →

Top score 85.4% by GPT-6 Astra - 1 model reporting (1 frontier)

Top models

1
SEC-Bench ProBar chart with 1 bar. Highest value: GPT-6 Astra at 85.4.
1 model

FAQ

What is SEC-Bench Pro?
Agentic cybersecurity benchmark covering reverse engineering and security analysis tasks.
What is the current top score on SEC-Bench Pro?
The top reported score is 85.4% by GPT-6 Astra, across 1 model reporting (1 from frontier labs).