Benchmark
Sayf CTI-Bench MCQ accuracy
Higher scores are better.
Benchmark sourceScores from Sayf cybersecurity evaluations data docs.
- Models measured
- 6
- Scores recorded
- 6
- Sources
- 1
- IQ weight
- 0.98
Score distribution
- 0.742
- Mean
- 0.744
- Median
- 0.06
- Std dev
- 0.693
- 10th percentile
- 0.841
- 90th percentile
- 0.656
- Min
This benchmark has a weight of 0.98 for chat models, measured over 6 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.
Leaderboard
| # | Model | Score |
|---|---|---|
| 1 | Claude Sonnet 4.6mcq· Sayf cybersecurity evaluations data docs | 0.841 |
| 2 | GPT 5.4mcq· Sayf cybersecurity evaluations data docs | 0.785 |
| 3 | Gemma 4 31B ITmcq· Sayf cybersecurity evaluations data docs | 0.744 |
| 4 | Qwen3.6 35B A3Bmcq· Sayf cybersecurity evaluations data docs | 0.73 |
| 5 | GPT OSS 20bmcq· Sayf cybersecurity evaluations data docs | 0.693 |
| 6 | Llama 3.3 70B Instructmcq· Sayf cybersecurity evaluations data docs | 0.656 |
Data credits: Sayf-eval. Copyright (c) 2026 Qatar Computing Research Institute (QCRI). MIT License. (licence)