Long Document Analysis Recommendation: Claude 3.5 Sonnet
Long Document Analysis → Claude 3.5 Sonnet
Aperçu du cas d’usage
Claude 3.5 Sonnet by Anthropic is a recommended option for the Long Document Analysis use case.
Modèles recommandés
| Rang | Fournisseur | Fenêtre de contexte | Score |
|---|
| #1 | Anthropic | 200K | 37/100 |
| Benchmark | Score |
|---|
| MMLU | 88.7 |
| HUMANEVAL | 92.0 |
| GSM8K | 96.4 |
| MATH | 71.1 |
| BBH | 84.5 |
Forces
- MMLU score 88.7, strong knowledge reasoning.
- HumanEval 92.0, excellent code generation.
- GSM8K 96.4, robust math reasoning.
- 上下文窗口 200K，支持长文本。
Exigences
- 需要 anthropic 的 API 密钥
- 输入长度须在 200K 上下文窗口内