GPT-6.1 Sol
工具调用推理缓存结构化输出
官方价格
$2.00 / $10.00
输入 / 输出 · OpenAI
最低输入价
$0.2545
Aiberm
上下文
1.1M
比较服务商
每个点是一家服务商。从左到右是每百万 token 的输入价,从下到上是输出价。虚线是官方价格。又左又下的点两边都更便宜。 离其余报价太远的价格没有画在这张图上。
基础价格
6最新价格检查于 Nov 18, 2026。
服务商价格
20金额均以美元计价。悬停或点击带下划线的金额,可查看价格换算明细。
DR 是服务商网站的 Ahrefs 域名评级,范围 0 到 100。域名年龄是该网站的注册时间。这些只说明公开痕迹,不能说明服务商是否会持续运营。域名评级低于 10 的报价不列出。
未验证报价(6)
金额均以美元计价。悬停或点击带下划线的金额,可查看价格换算明细。
DR 是服务商网站的 Ahrefs 域名评级,范围 0 到 100。域名年龄是该网站的注册时间。这些只说明公开痕迹,不能说明服务商是否会持续运营。域名评级低于 10 的报价不列出。
GPT-6.1 Sol · 已公布价格
| Standard tier | Input | Output | Cache read |
|---|---|---|---|
| Up to 272K | $2 | $10 | $0.10 |
| Over 272K | $4 | $15 | $0.20 |
GPT-6.1 Sol · 权重许可证
| 模型权重 | 协议 |
|---|---|
| 闭源 / 专有 | 闭源 / 专有 |
缓存与处理模式
上下文与限制
关于 GPT-6.1 Sol
Yes. Above 272,000 input tokens, OpenAI prices the full request at 2 times the input and cache rates and 1.5 times the output rate. Standard rates below that threshold are $2 input, $10 output, and $0.10 cached input per million tokens. The long-context Standard output rate is $15, not $10.
Fast costs twice Standard, and Ultrafast costs six times Standard. Batch and Flex cost 50 percent less than Standard. Regional processing adds 10 percent where it is available. Apply the mode multiplier to the short-context or long-context row that matches the prompt, not to a mix of the two.
On short-context Standard, 1 million input tokens and 100,000 output tokens cost $3 before cache and regional charges. The same shape above 272,000 input tokens costs $5.50. Reading 1 million cached input tokens costs $0.10 on the short-context Standard row, after a cache write at $2.50 per million. These are aggregate workload totals across requests, with every individual prompt within the short-tier threshold; they are not one oversized prompt.
Short-context Fast is $4 input and $20 output per million tokens. Ultrafast is $12 and $60. Batch and Flex are $1 and $5. Above 272,000 input tokens, apply the 2 times input and 1.5 times output multipliers first, then the mode multiplier. A 10 percent regional premium, where available, comes after that.
For a GPT 6.1 Sol cost estimate, count input, cached input, and output separately and multiply each by its applicable per-million-token rate. Use the long-context tier for the full request when the prompt exceeds 272K tokens; check the selected processing mode before comparing quotes.
来源与注脚
- [1]来源核验日期:
- [2]
输入、输出和缓存读取的单位是每百万 token 的美元价格。速度是每秒输出 token 数。可用率是来源报告的在线比例。













