CyberCertBench - Fortinet NSE: leaderboard
Metric: Accuracy (%) on 141 Fortinet NSE certification questions (procedural, vendor-specific configuration of the Fortinet security ecosystem): 5-shot multiple-choice question answering at temperature 0, scored by exact accuracy (correct only when the model selects all correct options and no incorrect one), mean accuracy (%) as printed in the appendix; questions scraped from public certification question banks and checked against authoritative documentation; higher is better. Source: arxiv.org. Saturation forecast: Around 2029. 45 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Sonnet 4.5 | 62.4 |
| 2 | DeepSeek V3 (0324) | 62.4 |
| 3 | Gemini 2.5 Pro (Preview 05-06) | 61.7 |
| 4 | Kimi K2 | 61 |
| 5 | Claude 3.5 Sonnet | 59.6 |
| 6 | GPT-4o | 58.2 |
| 7 | GPT-4.1 | 58.2 |
| 8 | Gemini 2.0 Flash | 56.7 |
| 9 | Claude Haiku 4.5 | 56 |
| 10 | Llama 4 Maverick | 56 |
| 11 | GPT-5 Mini | 55.3 |
| 12 | Mistral Medium 3.1 | 55.3 |
| 13 | Mixtral 8x22B | 55.3 |
| 14 | GPT-OSS-20B | 53.9 |
| 15 | Qwen 2.5 72B | 53.9 |
Interactive version: theaggregate.ai/benchmark?slug=cybercertbench-fortinet-nse · How It Works · Data refreshed daily, snapshot 2026-10-07.