Benchmark
PGG-Bench (Lech Mazur)
Scores are percentages.
Benchmark sourceScores from Lech Mazur suite data.
- Models measured
- 17
- Scores recorded
- 18
- Sources
- 1
- IQ weight
- 1.17
Score distribution
- 10%
- Mean
- 10%
- Median
- 2%
- Std dev
- 8.3%
- 10th percentile
- 11.9%
- 90th percentile
- 5.1%
- Min
This benchmark has a weight of 1.17 for chat models, measured over 17 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 | Gemini 2.5 Pro· Lech Mazur suite data | 14.3% |
| 2 | o1medium reasoning· Lech Mazur suite data | 13.4% |
| 3 | Mistral Large 2· Lech Mazur suite data | 11.9% |
| 4 | gemini-2.0-pro-exp-02-05· Lech Mazur suite data | 11.1% |
| 5 | o3 Minimedium reasoning· Lech Mazur suite data | 11.1% |
| 6 | DeepSeek V3· Lech Mazur suite data | 10.9% |
| 7 | GPT-4.5 Preview· Lech Mazur suite data | 10.8% |
| 8 | Llama 3.3 70B Instruct· Lech Mazur suite data | 10.2% |
| 9 | Claude Sonnet 3.7thinking, 16k· Lech Mazur suite data | 10% |
| 10 | GPT 4o· Lech Mazur suite data | 9.9% |
| 11 | DeepSeek R1· Lech Mazur suite data | 9.7% |
| 12 | Qwen2.5-Max· Lech Mazur suite data | 9.5% |
| 13 | Claude Sonnet 3.7· Lech Mazur suite data | 9.3% |
| 14 | QwQ 32B16k· Lech Mazur suite data | 8.6% |
| 15 | Llama 3.1 405B· Lech Mazur suite data | 8.5% |
| 16 | Gemini 2.0 Flash Experimentalthinking· Lech Mazur suite data | 8.3% |
| 17 | Claude 3.5 Sonnet· Lech Mazur suite data | 8.1% |
| 18 | Gemini 2.0 Flash· Lech Mazur suite data | 5.1% |