Benchmark
Multi-Agent Step Race (Lech Mazur)
Agentic benchmarks measure how well models carry out tasks that take several steps. Scores are percentages.
- Models measured
- 52
- Scores recorded
- 59
- Sources
- 1
- IQ weight
- 1.86
Score distribution
- 2.3%
- Mean
- 2%
- Median
- 1.3%
- Std dev
- 1%
- 10th percentile
- 3.8%
- 90th percentile
- 0.74%
- Min
This benchmark has a weight of 1.86 for chat models, measured over 52 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 |
|---|---|---|
| 51 | Mistral Large 2· Lech Mazur suite data | 1.1% |
| 52 | Gemini 2.0 Flash Experimental· Lech Mazur suite data | 1% |
| 53 | Gemini 1.5 Pro· Lech Mazur suite data | 1% |
| 54 | GPT-4.5 Preview· Lech Mazur suite data | 1% |
| 55 | Gemini 1.5 Flash· Lech Mazur suite data | 0.92% |
| 56 | Llama 3.1 405B· Lech Mazur suite data | 0.91% |
| 57 | GPT-4o mini· Lech Mazur suite data | 0.87% |
| 58 | Claude Haiku 3.5· Lech Mazur suite data | 0.78% |
| 59 | gemma2-27b· Lech Mazur suite data | 0.74% |
Showing 51–59 of 59 results.