Logical Reasoning Recommendation: GPT-4 32K
Logical Reasoning → GPT-4 32K
Anwendungsfallübersicht
GPT-4 32K by Openai is a recommended option for the Logical Reasoning use case.
Empfohlene Modelle
| Rang | Anbieter | Kontextfenster | Score |
|---|
| #8 | Openai | 32K | 56/100 |
Leistungsanalyse
Benchmark-Leistung
| Benchmark | Ergebnis |
|---|
| MMLU | 80.8 |
| HUMANEVAL | 76.8 |
| GSM8K | 81.6 |
| MATH | 40.1 |
| BBH | 79.2 |
| GPQA | 40.9 |
Stärken
- MMLU score 80.8, strong knowledge reasoning.
- 上下文窗口 32K，支持长文本。
Anforderungen
- 需要 openai 的 API 密钥
- 输入长度须在 32K 上下文窗口内