The latest GPT-4 model with improved performance, updated knowledge, and enhanced capabilities. It offers faster response times and more affordable pricing compared to previous versions.
发布日期2024年4月9日
参数规模—
上下文长度128K
许可证Proprietary
知识截止2023年12月31日
Benchmarks
评测成绩
| 评测基准 | 类别 | 分数 | 来源 |
|---|---|---|---|
| 安全防御榜·Toxicity | security | 96.9 | — |
| 安全防御榜·Human Rights | security | 94.8 | — |
| 安全防御榜·Malicious Use | security | 94.4 | — |
| 安全防御榜·Info Safety | security | 94.2 | — |
| 安全防御榜 | security | 93.8 | — |
| 安全防御榜·Bias | security | 93.6 | — |
| 安全防御榜·Child Safety | security | 91.5 | — |
| 安全防御榜·Socioeconomic | security | 91.4 | — |
| MGSM | math reasoning | 88.5 | 来源 |
| HumanEval | reasoning code | 87.1 | 来源 |
| MMLU | general reasoning language math | 86.5 | 来源 |
| DROP | reasoning math | 86.0 | 来源 |
| 价值对齐榜·Fairness | security | 72.9 | — |
| MATH | math reasoning | 72.6 | 来源 |
| 价值对齐榜·Legality | security | 72.5 | — |
| 价值对齐榜·Data Protection | security | 72.3 | — |
| 价值对齐榜 | security | 67.6 | — |
| 价值对齐榜·Safety | security | 65.1 | — |
| 价值对齐榜·Morality | security | 55.4 | — |
| GPQA | reasoning general | 48.0 | 来源 |
Pricing
API 价格对比
| 服务商 | 输入价 | 输出价 | 上下文 | 吞吐(tok/s) | 延迟(s) | 函数调用 | 代码执行 | 联网搜索 |
|---|---|---|---|---|---|---|---|---|
| Azure | $10.00 | $30.00 | 128K | — | — | ✓ | ✗ | ✗ |
| Azure Cognitive Services | $10.00 | $30.00 | 128K | — | — | ✓ | ✗ | ✗ |
| LLM Gateway | $10.00 | $30.00 | 128K | — | — | ✓ | ✗ | ✗ |
| OpenAI | $10.00 | $30.00 | 128K | — | — | ✓ | ✗ | ✗ |
价格单位：美元/百万 token，数据来自社区整理，仅供参考。