Llama 2-13B vs GPT-3.5 Turbo (older v0613)
Côte à côte. Chaque métrique. Chaque benchmark.
| Type | Llama 2-13B | GPT-3.5 Turbo (older v0613) |
|---|---|---|
| Provider | ||
| score moyen | 42.5 | 45.8 |
| Prix d'entrée | - | $1.00 |
| Prix de sortie | - | $2.00 |
| Fenêtre de contexte | - | 4K tokens (~2 books) |
| Sorti le | 2024-01-01 | 2024-01-25 |
| Code source ouvert | Open Source | Proprietary |
Scores de benchmark
10 benchmarks · Llama 2-13B: 0, GPT-3.5 Turbo (older v0613): 10
| Benchmark | Catégorie | Llama 2-13B | GPT-3.5 Turbo (older v0613) |
|---|---|---|---|
| ARC AI2 | knowledge | 47.1 | 83.2 |
| BBH | reasoning | 44.3 | 48.8 |
| CSQA2 | knowledge | 0.1 | 14.0 |
| GPQA diamond | knowledge | 1.8 | 2.9 |
| GSM8K | math | 36.9 | 57.8 |
| MATH level 5 | math | 3.3 | 11.6 |
| MMLU | knowledge | 40.8 | 56.4 |
| OpenBookQA | knowledge | 42.7 | 81.3 |
| TriviaQA | knowledge | 79.6 | 85.8 |
| Winogrande | knowledge | 45.6 | 63.2 |