SmolLM-360M: benchmark results
Hugging Face's SmolLM-360M, the middle size of three SmolLM models (135M, 360M, 1.7B), trained on the Cosmo-Corpus dataset combining synthetic textbooks, Python code and educational web text. Provider: HuggingFace. Released 2024-07-16. Access: Open.
Aggregate Rating 1213 ± 39, rank #1579 of 1661 rated models, from 39 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AI Energy Score (Text Generation) | 4 | Energy Score (1-5) | 47.5 |
| Open LLM Leaderboard - MuSR | 40.18 | Score | 43.3 |
| Open PL LLM - CBD (multiple choice, 5-shot) | 16.26 | Macro F1 (%) | 26.9 |
| Open PL LLM - PSC (generative, 5-shot) | 74.56 | Binary F1 (%) | 23.3 |
| Open LLM Leaderboard - GPQA | 26.76 | Score | 22 |
| Open Portuguese LLM - OAB Exams | 26.97 | Accuracy (%) | 20.9 |
| Open Portuguese LLM - BLUEX | 25.31 | Accuracy (%) | 20.8 |
| Open PL LLM - CBD (generative, 5-shot) | 15.74 | Macro F1 (%) | 17.9 |
| Open Portuguese LLM - FaQuAD NLI | 43.97 | Macro F1 (%) | 17.1 |
| Open LLM Leaderboard - IFEval | 21.34 | Score | 13.8 |
| Open PL LLM - PPC (generative, 5-shot) | 37.9 | Accuracy (%) | 13.8 |
| Open PL LLM - DYK (generative, 5-shot) | 23.11 | Binary F1 (%) | 12.6 |
Interactive version: theaggregate.ai/model?slug=smollm-360m · How It Works · Data refreshed daily, snapshot 2026-10-23.