Model overview
纯官方GPT4 32K系列，支持更大的上下文窗口，适用于复杂的对话场景。将于 2025-06-06 废弃，建议使用 gpt-4o
The summary below is compiled from sources by an AI model, not our own testing. Verify selection decisions against original sources and your workload.
Key features
- 支持32K上下文窗口，适合处理长文本和复杂对话
- 为官方GPT-4系列模型，具备GPT-4的基础能力
- 专为高要求对话场景设计
Limitations & considerations
- 将于2025-06-06正式废弃
- 官方已明确建议迁移至gpt-4o
Source prices & cost analysis
Prices vary by channel, region and configuration. Channel quotes are listed independently and are not treated as a single official provider price.
| Source / channel | Input | Output | Billing unit | Collected at |
|---|---|---|---|---|
| Apilio · AZ-优质 ↗ | Not verified | Not verified | 渠道倍率待核验 | 2026/11/2 06:25:16 |
| Apilio · default ↗ | Not verified | Not verified | 渠道倍率待核验 | 2026/11/2 06:25:16 |
| Apilio · origin ↗ | Not verified | Not verified | 渠道倍率待核验 | 2026/11/2 06:25:16 |
| Apilio · ssvip ↗ | Not verified | Not verified | 渠道倍率待核验 | 2026/11/2 06:25:16 |
| Apilio · svip ↗ | Not verified | Not verified | 渠道倍率待核验 | 2026/11/2 06:25:16 |
| Apilio · vip ↗ | Not verified | Not verified | 渠道倍率待核验 | 2026/11/2 06:25:16 |
| Apilio · vvip ↗ | Not verified | Not verified | 渠道倍率待核验 | 2026/11/2 06:25:16 |
Estimate monthly base token costs
Estimates exclude cache discounts, tiered pricing, multimodal surcharges, web search, taxes and additional reasoning output.
Public benchmarks & run configuration
暂无已采集的可核验公开评分。模型未上榜不代表能力较低。
View full indexed rankings →Frequently asked questions
gpt-4-32k 是否仍是当前可用的模型？
是的，截至2025-06-06前仍可用，但已进入废弃倒计时。
为什么建议改用 gpt-4o？
因gpt-4o在响应速度、多模态能力及成本效率等方面具有综合优势，官方推荐替代。
废弃后还能继续调用 gpt-4-32k 吗？
废弃后API端点将停止服务，无法再成功调用。
Do open weights permit free commercial use?
Not necessarily. Check the version license, commercial conditions and usage policies. Self-hosting incurs compute and maintenance costs.
Are these real-time billing prices?
These are timestamped source snapshots, not live billing prices. Check the channel for current versions, regions, discounts and billing methods.
Sources & traceability
许可证：来源未披露，须进一步核验。模型介绍与第三方评测均以原来源的具体版本和方法为准。