SmolLM2-1.7B-Instruct: benchmark results
An instruction-tuned checkpoint for on-device text rewriting, summarization and function calling. Its post-training combines supervised examples with preference optimization. Provider: HuggingFace. Released 2024-10-31. Access: Open.
Aggregate Rating 1233 ± 13, rank #1562 of 1656 rated models, from 250 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| LA Leaderboard - MGSM Spanish | 4.4 | Exact Match (%) | 82.3 |
| LA Leaderboard - QNLI Basque | 54.62 | Accuracy (%) | 81.6 |
| LA Leaderboard - RagQuAS | 80.05 | Semantic Answer Similarity (x100) | 80.9 |
| Open CoT - LSAT Analytical Reasoning | 6.09 | CoT Gain (%) | 76.7 |
| LA Leaderboard - ClinTreatES | 60.81 | Semantic Answer Similarity (x100) | 76.1 |
| MEDIC Benchmark | 72.15 | MEDIC Public Table Average | 75.8 |
| LA Leaderboard - caBREU Catalan | 12.67 | BLEU | 70.6 |
| LA Leaderboard - Summarization Galician | 5.24 | BLEU | 68.7 |
| Open LLM Leaderboard - IFEval | 53.68 | Score | 64.3 |
| LA Leaderboard - NoticIA | 12.64 | ROUGE-1 (%) | 61.5 |
| LA Leaderboard - MGSM Catalan | 2.8 | Exact Match (%) | 56.9 |
| SLM-RAG Arena | 1493.2 | Elo Rating | 56.2 |
Interactive version: theaggregate.ai/model?slug=smollm2-1-7b-instruct · How It Works · Data refreshed daily, snapshot 2026-10-20.