Qwen2.5 0.5B Instruct
Código abiertopor Alibaba · Publicado el 2024-09-16
10.1
puntuación promedio
N/A
Precio de entrada
N/A
Precio de salida
N/A
Ventana de contexto
text-generation
Tipo
Tested on 6 benchmarks with 10.1% average. Top scores: IFEval (31.5%), MATH Level 5 (10.3%), BBH (HuggingFace) (8.2%).
Puntuaciones de benchmark
| Benchmark | Categoría | Puntuación | Bar |
|---|---|---|---|
| IFEval | language | 31.5 | |
| MATH Level 5 | math | 10.3 | |
| BBH (HuggingFace) | general | 8.2 | |
| MMLU-PRO | knowledge | 8.0 | |
| MUSR | reasoning | 1.4 | |
| GPQA | knowledge | 1.2 |
Modelos similares
Gemma 2 2bGoogle DeepMind
10.4
DeepSeek R1 Distill Qwen 1.5BDeepSeek
10.4
L
Vicuna 7b V1.5LMSYS
10.9
Llama 3.2 3B Instruct (free)Meta
8.7