Multi-step Planning Recommendation: Llama 3.1 405B
Multi-step Planning → Llama 3.1 405B
Kullanım Senaryosu Genel Bakışı
Llama 3.1 405B by Meta is a recommended option for the Multi-step Planning use case.
Önerilen Modeller
| Sıralama | Satıcı | Bağlam Penceresi | Score |
|---|
| #6 | Meta | 128K | 37/100 |
| Benchmark | Puan |
|---|
| MMLU | 88.6 |
| HUMANEVAL | 89.0 |
| GSM8K | 89.2 |
| MATH | 73.8 |
| BBH | 82.9 |
Güçlü Yönler
- MMLU score 88.6, strong knowledge reasoning.
- HumanEval 89.0, excellent code generation.
- GSM8K 89.2, robust math reasoning.
Gereksinimler
- 需要 meta 的 API 密钥
- 输入长度须在 128K 上下文窗口内