Benchmark
SimpleBench
Scores are percentages.
All reasoning benchmarksScores from Epoch AI Benchmarking Hub data.
- Models measured
- 105
- Scores recorded
- 114
- Sources
- 1
- IQ weight
- 1.75
Score distribution
- 54.6%
- Mean
- 56.7%
- Median
- 20.9%
- Std dev
- 23.5%
- 10th percentile
- 81.9%
- 90th percentile
- 10.7%
- Min
This benchmark has a weight of 1.75 for chat models, measured over 105 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 |
|---|---|---|
| 101 | GPT-4 Turbo· Epoch AI Benchmarking Hub data | 25.1% |
| 102 | MiniMax M2· Epoch AI Benchmarking Hub data | 25% |
| 103 | Claude 3 Opus· Epoch AI Benchmarking Hub data | 23.5% |
| 104 | Llama 3.1 405B Instruct· Epoch AI Benchmarking Hub data | 23% |
| 105 | o3 Minihigh· Epoch AI Benchmarking Hub data | 22.8% |
| 106 | Grok 2· Epoch AI Benchmarking Hub data | 22.7% |
| 107 | Mistral Large2407· Epoch AI Benchmarking Hub data | 22.5% |
| 108 | GPT OSS 120B· Epoch AI Benchmarking Hub data | 22.1% |
| 109 | Mistral Large 3· Epoch AI Benchmarking Hub data | 20.4% |
| 110 | Llama 3.3 70B Instruct· Epoch AI Benchmarking Hub data | 19.9% |
| 111 | Gemini 2.0 Flash Experimental· Epoch AI Benchmarking Hub data | 18.9% |
| 112 | O1 Minimedium· Epoch AI Benchmarking Hub data | 18.1% |
| 113 | GPT 4o· Epoch AI Benchmarking Hub data | 17.8% |
| 114 | GPT-4o mini· Epoch AI Benchmarking Hub data | 10.7% |
Showing 101–114 of 114 results.
Data credits: Epoch AI Benchmarking Hub, CC BY 4.0 (licence)