A compact but capable model that punches above its weight for its size class. Qwen2.5 3B handles instruction-following and multilingual tasks — particularly Chinese and English — with surprising coherence, though it naturally hits limits on complex multi-step reasoning or deep technical problems that larger models handle more gracefully. It runs efficiently on modest hardware, making it practical for edge deployments or resource-constrained environments.
| Benchmark | Score | Type | Recorded |
|---|---|---|---|
| IFEval | 26.9 | accuracy | 4d ago |