Qwen2.5 Coder 14B — Benchmarks
Benchmark scores for Qwen2.5 Coder 14B aggregated from public leaderboards, with how it ranks among open models. See hardware requirements for what you need to run it.
Overall rank: #9 of 123 open modelscomposite 70.4/100 across 5 benchmarks in 3 categories · methodology
Knowledge
| Benchmark | Score | Open rank | All models |
|---|---|---|---|
| HellaSwag | 80.2% | #15 / 42 | #28 / 76 |
| ARC Challenge | 66.0% | #20 / 51 | #29 / 77 |
| MMLU | 75.2% | #19 / 77 | #44 / 136 |
Math
| Benchmark | Score | Open rank | All models |
|---|---|---|---|
| GSM8K | 88.7% | #6 / 60 | #9 / 93 |
Reasoning
| Benchmark | Score | Open rank | All models |
|---|---|---|---|
| WinoGrande | 76.8% | #15 / 46 | #32 / 80 |
Scores aggregated from public benchmark sources (each linked from the benchmark pages). llmrun does not run these benchmarks.