SmolLM2 135M Instruct vs Qwen2 0.5B Instruct
Lado a lado. Cada métrica. Cada benchmark.
| Tipo | SmolLM2 135M Instruct | Qwen2 0.5B Instruct |
|---|---|---|
| Provider | HF Hugging Face TB |  Alibaba |
| puntuación promedio | 15.0 | 6.6 |
| Precio de entrada | - | - |
| Precio de salida | - | - |
| Ventana de contexto | - | - |
| Publicado el | 2024-10-31 | 2024-06-03 |
| Código abierto | Open Source | Open Source |
Puntuaciones de benchmark
6 benchmarks · SmolLM2 135M Instruct: 2, Qwen2 0.5B Instruct: 3
| Benchmark | Categoría | SmolLM2 135M Instruct | Qwen2 0.5B Instruct |
|---|---|---|---|
| BBH (HuggingFace) | general | 4.7 | 5.9 |
| GPQA | knowledge | 0.0 | 0.0 |
| IFEval | language | 28.8 | 22.5 |
| MATH Level 5 | math | 0.3 | 2.9 |
| MMLU-PRO | knowledge | 1.3 | 5.9 |
| MUSR | reasoning | 3.7 | 2.4 |