Benchmark Performance
| Benchmark | Score | Unit | Evaluated At | Notes | Source |
|---|
| MMLU | 61.1 | % | 2023-03-01 | 5-shot | view |
| HUMANEVAL | 51.3 | pass@1 | 2023-03-01 | — | view |
| GSM8K | 34.3 | % | 2023-03-01 | 0-shot CoT | view |
| MATH | 17.8 | % | 2023-03-01 | 0-shot CoT | view |
| BBH | 48.2 | % | 2023-03-01 | 3-shot CoT | view |
| GPQA | 22.4 | % | 2023-03-01 | 0-shot | view |
| IFEVAL | 57.4 | % | 2023-03-01 | prompt_strict | view |
| ARC | 86.8 | % | 2023-03-01 | challenge | view |
| MUSR | 29.2 | % | 2023-03-01 | 0-shot | view |
| WINOGRANDE | 76.7 | % | 2023-03-01 | 0-shot | view |
Pricing
| Tier | Price | Currency |
|---|
| Input | $0.5 / Mtok | USD |
| Output | $1.5 / Mtok | USD |
| Cache Read | $0 / Mtok | USD |
| Cache Write | $0 / Mtok | USD |
Source:
https://openai.com/api/pricing/
· as of 2023-03-01
Compliance
- Data Residency: US
- SOC2: ✓
- HIPAA: ✓
- GDPR: ✓
- ISO 27001: ✓
GPT-3.5 Turbo
Visão geral do modelo
OpenAI GPT-3.5 Turbo 经济型模型, 16K 上下文, 速度快价格低, 适合对话与通用任务。
Especificações principais
| Fornecedor | Versão | Data de lançamento | Janela de contexto | Modalidades de entrada | Modalidades de saída | Licença |
|---|
| Openai | 3.5-turbo | 2023-03-01 | 16K | text | text | Proprietary |
Desempenho em benchmarks
| Benchmark | Pontuação | Unidade | Notas |
|---|
| MMLU (Massive Multitask Language Understanding) | 61.1 | % | 5-shot |
| HumanEval | 51.3 | pass@1 | — |
| GSM8K (Grade School Math 8K) | 34.3 | % | 0-shot CoT |
| MATH | 17.8 | % | 0-shot CoT |
| BBH (BIG-Bench Hard) | 48.2 | % | 3-shot CoT |
| GPQA | 22.4 | % | 0-shot |
| IFEval | 57.4 | % | prompt_strict |
| ARC | 86.8 | % | challenge |
| MUSR | 29.2 | % | 0-shot |
| WinoGrande | 76.7 | % | 0-shot |
Preços
| Entrada | Saída | Leitura de cache | Escrita de cache |
|---|
| — | — | — | — |
por milhão de tokens
Pontos fortes
Pontos fracos
- 闭源专有模型，不支持自托管。
- 上下文窗口 16K 偏小。
Casos de uso
Referências