Benchmark
CyBench Subtasks
Coding benchmarks compare models on programming tasks. Scores are percentages.
- Models measured
- 7
- Scores recorded
- 7
- Sources
- 1
- IQ weight
- 0.58
Score distribution
- 29.1%
- Mean
- 28.7%
- Median
- 12.9%
- Std dev
- 15.2%
- 10th percentile
- 43.9%
- 90th percentile
- 11.7%
- Min
This benchmark has a weight of 0.58 for chat models, measured over 7 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 | O1 Preview· CyBench official data | 46.8% |
| 2 | Claude 3.5 Sonnet· CyBench official data | 43.9% |
| 3 | Claude 3 Opus· CyBench official data | 36.8% |
| 4 | GPT 4o· CyBench official data | 28.7% |
| 5 | Llama 3.1 405B Instruct· CyBench official data | 20.5% |
| 6 | Mixtral 8x22B Instruct· CyBench official data | 15.2% |
| 7 | Gemini 1.5 Pro· CyBench official data | 11.7% |
Data credits: CyBench — cite Zhang et al. 2025