Qwen/Qwen2-57B-A14B
| benchmark | category | score | context | conf | evidence |
|---|---|---|---|---|---|
| HellaSwag | knowledge | 87.1 | HellaSwag | medium | | HellaSwag | **87.1** | 86.5 | 85.9 | 85.0 | 85.2 | |
| CMMLU | knowledge | 84.8 | CMMLU | medium | | CMMLU | - | - | 84.8 | 82.3 | **88.5** | |
| C-Eval | knowledge | 83.5 | C-Eval | medium | | C-Eval | - | - | - | 83.5 | **87.7** | |
| WinoGrande | knowledge | 82.5 | Winogrande | medium | | Winogrande | 82.5 | 81.9 | **84.9** | 81.5 | 79.5 | |
| MMLU | knowledge | 67.4 | MMLU | medium | | MMLU | 67.4 | 71.8 | **77.1** | 74.3 | 76.5 | |
| MBPP | coding | 63.9 | MBPP | medium | | MBPP | - | 63.9 | 65.5 | 64.2 | **71.9** | |
| GSM8K | math | 59.9 | GSM8K | medium | | GSM8K | 59.9 | 62.5 | **82.7** | 76.8 | 80.7 | |
| TruthfulQA | safety | 46.4 | TruthfulQA | medium | | TruthfulQA | 46.4 | 51.1 | 53.9 | 57.4 | **57.7** | |
| EvalPlus | coding | 46.4 | EvalPlus | medium | | EvalPlus | - | 46.4 | 51.9 | 50.4 | **57.2** | |
| BBH | reasoning | 45.4 | BBH | medium | | BBH | 45.4 | 50.3 | **76.4** | 66.8 | 67.0 | |
| MMLU-Pro | knowledge | 41.0 | MMLU-Pro | medium | | MMLU-Pro | - | 41.0 | **48.3** | 44.0 | 43.0 | |
| MATH | math | 30.8 | MATH | medium | | MATH | - | 30.8 | 41.7 | 36.1 | **43.0** | |
| HumanEval | coding | 29.3 | HumanEval | medium | | HumanEval | 29.3 | 37.2 | 46.3 | 43.3 | **53.0** | |
| GPQA | reasoning | 29.2 | GPQA | medium | | GPQA | - | 29.2 | - | 30.8 | **34.3** | |
none scraped.