Qwen2.5 7B Instruct
Qwen2.5-7B-Instruct is an instruction-tuned 7B parameter language model that excels at following instructions, generating long texts (over 8K tokens), understanding structured data, and generating structured outputs like JSON. The model features enhanced capabilities in mathematics, coding, and multilingual support across 29+ languages including Chinese, English, French, Spanish, and more.
Benchmark results
| Benchmark | Score | Tags | Source |
|---|---|---|---|
| AlignBench | 73.3% | self-reported llm-stats | link → |
| Arena Hard | 52.0% | self-reported llm-stats | link → |
| GPQA | 36.4% | self-reported llm-stats | link → |
| GSM8k | 91.6% | self-reported llm-stats | link → |
| HumanEval | 84.8% | self-reported llm-stats | link → |
| IFEval | 71.2% | self-reported llm-stats | link → |
| LiveBench | 35.9% | self-reported llm-stats | link → |
| LiveCodeBench | 28.7% | self-reported llm-stats | link → |
| MATH | 75.5% | self-reported llm-stats | link → |
| MBPP | 79.2% | self-reported llm-stats | link → |
| MMLU-Pro | 56.3% | self-reported llm-stats | link → |
| MMLU-Redux | 75.4% | self-reported llm-stats | link → |
| MT-Bench | 87.5 | self-reported llm-stats | link → |
| MultiPL-E | 70.4% | self-reported llm-stats | link → |