Benchmark
Sayf SEvenLLM-Bench (English) accuracy
Higher scores are better.
Benchmark sourceScores from Sayf cybersecurity evaluations data docs.
- Models measured
- 6
- Scores recorded
- 6
- Sources
- 1
- IQ weight
- 0.89
Score distribution
- 0.871
- Mean
- 0.869
- Median
- 0.034
- Std dev
- 0.854
- 10th percentile
- 0.928
- 90th percentile
- 0.823
- Min
This benchmark has a weight of 0.89 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 4.6sevenllm· Sayf cybersecurity evaluations data docs | 0.928 |
| 2 | GPT 5.4sevenllm· Sayf cybersecurity evaluations data docs | 0.898 |
| 3 | Gemma 4 31B ITsevenllm· Sayf cybersecurity evaluations data docs | 0.869 |
| 4 | GPT OSS 20bsevenllm· Sayf cybersecurity evaluations data docs | 0.855 |
| 5 | Llama 3.3 70B Instructsevenllm· Sayf cybersecurity evaluations data docs | 0.854 |
| 6 | Qwen3.6 35B A3Bsevenllm· Sayf cybersecurity evaluations data docs | 0.823 |
Data credits: Sayf-eval. Copyright (c) 2026 Qatar Computing Research Institute (QCRI). MIT License. (licence)