A compact, efficient model that punches above its weight for its size class. Qwen2 1.5B handles everyday language tasks — summarization, Q&A, basic reasoning — with reasonable coherence, though it naturally hits walls on complex multi-step problems that larger models handle more gracefully. Its open-weight Apache license makes it easy to run locally or fine-tune.
| Benchmark | Score | Type | Recorded |
|---|---|---|---|
| IFEval | 21.1 | accuracy | 4d ago |