RedSage-Bench: leaderboard

Cybersecurity benchmark with 30K multiple-choice and 240 open-ended QA items covering knowledge, offensive skills, and tool expertise.

Metric: Macro Accuracy (self-reported). Source: benchmarklist.com. Status: saturated. 17 models tracked.

Top models

#ModelScore
1GPT-588.68
2Qwen 3 32B85.4
3Qwen3-8B-Base84.24
4Qwen 3 8B81.85
5Foundation-Sec-8B78.51
6Llama 3.1 8B78.02
7Llama 3.1 8B Instruct77.05
8Foundation-Sec-8B-Instruct76.12

Interactive version: theaggregate.ai/benchmark?slug=redsage-bench · How It Works · Data refreshed daily, snapshot 2026-09-05.