Site descriptions, notices, and reviews are shown in the language in which they were submitted.
运营稳定，自营号池，3g上下行高速物理机。欢迎尝试。
Price listing
GPT 5.6 Sol
4 public records are listed; open the full page to review details as needed.
Price listing
GPT 5.6 Terra
1 public records are listed; open the full page to review details as needed.
Price listing
GPT 5.5
1 public records are listed; open the full page to review details as needed.
Price listing
GPT Image 2
1 public records are listed; open the full page to review details as needed.
Current public prices are shown by model and channel. Token prices use CNY per 1M tokens unless another billing unit is shown.
| Channel | Provider model | Current price | Price trend | Uptime | Fake % | Latency |
|---|---|---|---|---|---|---|
| 临促分组 | gpt-5.6-sol | 输入 ¥0.65 / 输出 ¥3.9 / 缓存 ¥0.065 / 写入 ¥0.8125 | +333% | — | — | — |
| gpt 易燃易爆炸 | gpt-5.6-sol | 输入 ¥1.15 / 输出 ¥6.9 / 缓存 ¥0.115 / 写入 ¥1.44 | +15% | — | — | — |
| vip | gpt-5.6-sol | 输入 ¥1.15 / 输出 ¥6.9 / 缓存 ¥0.115 / 写入 ¥1.44 | +15% | — | — | — |
| Channel | Provider model | Current price | Price trend | Uptime | Fake % | Latency |
|---|---|---|---|---|---|---|
| 蒸馏分组 | gpt-5.6-terra | 输入 ¥2 / 输出 ¥12 / 缓存 ¥0.2 / 写入 ¥2.5 | 0% | — | — | — |
| Channel | Provider model | Current price | Price trend | Uptime | Fake % | Latency |
|---|---|---|---|---|---|---|
| 蒸馏分组 | gpt-5.5 | 输入 ¥5 / 输出 ¥30 / 缓存 ¥0.5 / 写入 ¥5 | 0% | — | — | — |
| Channel | Provider model | Current price | Price trend | Uptime | Fake % | Latency |
|---|---|---|---|---|---|---|
| 生图非4k | gpt-image-2 | ¥0.03 / 次 | - | — | — | — |
openai目前加强风控，只有gpt5.5故障率低。grok调价至0.05x给大家先应急使用。也可以加群820116443获取最新资讯，有需要退款的也可以联系我。
给大家解释一下整体使用中会有2-5%的概率出现慢首字，是因为调整了号池的重试和调度，缺点是2-5%的概率出现慢首字，但是优点是不掉缓存。我大量的进行了下测试，sub2和cpa的原版调度的“慢尾逃逸”有点太激进了，逃逸一次炸了缓存就是10倍的扣费，我感觉性价比不太高。
欢迎使用浮生云算 API 中转服务！ 主力推荐模型： GPT-5.6sol / grok4.6 遇到问题，请加 QQ 群：820116443 价格随着号池难易程度实时波动，最新价格会在Q群同步。 📢 2026-05-06 上线 image 子域：图像生成（gpt-image-2 等）请把 base url 改成 https://image.fushengyunsuan.cn/v1 ，TTFB <1s 不再卡顿；API key 不变，文本继续走主域。详见首页「API 信息」。
近期部分首字上涨的原理是最近openai风控加强，降低了很多账号的权重，导致扛不住并发，4以上并发就会频繁429，然后调整优化了一下重试的逻辑，和加大了一下会话粘性，这样子不会因为429导致任务中断或者因为报错引起会话逃逸导致缓存爆炸，目前是整体权衡了一下这样子收益比较大，大家可以参考这个思路改一下自己的号池。
近期蒸馏封号的概率非常高，所以蒸馏分组价格调整至1x同官key了，价格非常高，不太建议大家蒸馏使用。
5.0 / 5 from 2 approved reviews
稳定,定期有活动,相比用下来很不多的
第一条评论是自己写的，虽然有点不太好意思，但是还是真心很欢迎大家来尝试一下的。
Uptime is the share of recent probes where the service was available; a higher value generally indicates better stability.
Model consistency uses model pass results to show whether the endpoint behaves like the target model.
Pricing and latency come from public monitoring data and are intended for side-by-side gateway comparison.
Data is provided for reference. Third-party gateways may experience outages or balance loss, so test first and add only a small balance when needed.