SEC-Bench Pro
Security exploitation agent benchmark; DeepSeek-V4.1-Flash evaluated with the Claude Code harness per the official setup; Pass@1.
Category
General Agents
Max Score
100.0
Score Type
percent
Active
Yes
Model Scores
Scores for SEC-Bench Pro
| Model | Score | Date | Verified |
|---|---|---|---|
| DeepSeek-V4.1-Flash |
100.00%
|
10.09.2026 | Verified |