简单透明的按量付费
按各模型的官方原价、按 token 计费。无订阅、无最低消费、无隐藏费用。
官方原价
各模型价格与厂商公开的官方价格一致,以美元计价。
只为用量付费
按每次请求实际使用的输入、输出和缓存 tokens 计费。
一个余额,通用所有模型
充值一次,即可在目录中的所有模型上使用。
账单透明
控制台逐条列出每次请求的 tokens 与费用。
一次请求如何计费
费用 = 输入 tokens × 输入单价 + 输出 tokens × 输出单价 + 缓存 tokens × 缓存单价
单价以每百万 tokens 计。命中缓存的输入 tokens 按(低得多的)缓存读取价计费。部分模型按输入长度阶梯计价;图像与多媒体模型可能按次计费。
价格表
官方原价,单位美元;token 价格为每百万 tokens。- gpt-5.6-lunaOpenAI·对话·922K阶梯计价输入$0.2输出$1.2缓存读取$0.02
- gpt-5.6-terraOpenAI·对话·922K阶梯计价输入$2输出$12缓存读取$0.2
- gpt-5.6-solOpenAI·对话·922K阶梯计价输入$4输出$20缓存读取$0.4
- gpt-6-lunaOpenAI·对话·922K阶梯计价输入$0.1输出$0.5缓存读取$0.01
- gpt-6-solOpenAI·对话·922K阶梯计价输入$2输出$10缓存读取$0.2
- gpt-6-astraOpenAI·对话·922K阶梯计价输入$10输出$50缓存读取$1
- gpt-4.1OpenAI·对话·1.05M输入$2输出$8缓存读取$0.5
- gpt-4.1-miniOpenAI·对话·1.05M输入$0.4输出$1.6缓存读取$0.1
- gpt-4.1-nanoOpenAI·对话·1.05M输入$0.1输出$0.4缓存读取$0.025
- gpt-4oOpenAI·对话·128K输入$2.5输出$10缓存读取$1.25
- gpt-4o-miniOpenAI·对话·128K输入$0.15输出$0.6缓存读取$0.075
- gpt-5OpenAI·对话·272K输入$1.25输出$10缓存读取$0.125
- gpt-5-miniOpenAI·对话·272K输入$0.25输出$2缓存读取$0.025
- gpt-5-nanoOpenAI·对话·272K输入$0.05输出$0.4缓存读取$0.005
- gpt-5-proOpenAI·对话·400K输入$15输出$120
- gpt-5-search-apiOpenAI·对话·272K输入$1.25输出$10缓存读取$0.125
- gpt-5.1OpenAI·对话·272K输入$1.25输出$10缓存读取$0.125
- gpt-5.2OpenAI·对话·272K输入$1.75输出$14缓存读取$0.175
- gpt-5.2-proOpenAI·对话·272K输入$21输出$168
- gpt-5.3-codexOpenAI·对话·272K输入$1.75输出$14缓存读取$0.175
- gpt-5.4OpenAI·对话·1.05M阶梯计价输入$2.5输出$15缓存读取$0.25
- gpt-5.4-miniOpenAI·对话·272K输入$0.75输出$4.5缓存读取$0.075
- gpt-5.4-nanoOpenAI·对话·272K输入$0.2输出$1.25缓存读取$0.02
- gpt-5.4-proOpenAI·对话·1.05M阶梯计价输入$30输出$180
- gpt-5.5OpenAI·对话·1.05M阶梯计价输入$5输出$30缓存读取$0.5
- gpt-5.5-proOpenAI·对话·1.05M阶梯计价输入$30输出$180
- gpt-5.6OpenAI·对话·922K阶梯计价输入$4输出$20缓存读取$0.4
- gpt-6.1-solOpenAI·对话·922K阶梯计价输入$2输出$10缓存读取$0.1
- gpt-audioOpenAI·对话·128K输入$2.5输出$10
- gpt-audio-1.5OpenAI·对话·128K输入$2.5输出$10
- gpt-audio-miniOpenAI·对话·128K输入$0.6输出$2.4
- o1OpenAI·对话·200K输入$15输出$60缓存读取$7.5
- o1-proOpenAI·对话·200K输入$150输出$600
- o3OpenAI·对话·200K输入$2输出$8缓存读取$0.5
- o3-miniOpenAI·对话·200K输入$1.1输出$4.4缓存读取$0.55
- o3-proOpenAI·对话·200K输入$20输出$80
- o4-miniOpenAI·对话·200K输入$1.1输出$4.4缓存读取$0.275
- gpt-image-1OpenAI·图像输入$5输出—缓存读取$1.25
- gpt-image-1-miniOpenAI·图像输入$2输出—缓存读取$0.2
- gpt-image-1.5OpenAI·图像输入$5输出$10缓存读取$1.25
- gpt-image-2OpenAI·图像输入$5输出—缓存读取$1.25
- text-embedding-3-largeOpenAI·向量·8K输入$0.13输出—
- text-embedding-3-smallOpenAI·向量·8K输入$0.02输出—
- text-embedding-ada-002OpenAI·向量·8K输入$0.1输出—
- claude-fable-5Anthropic·对话·1M输入$10输出$50缓存读取$1
- claude-fable-5-1Anthropic·对话·1M输入$10输出$50缓存读取$0.25
- claude-haiku-4-5Anthropic·对话·200K输入$1输出$5缓存读取$0.1
- claude-mythos-5Anthropic·对话·1M输入$10输出$50缓存读取$1
- claude-mythos-5-1Anthropic·对话·1M输入$10输出$50缓存读取$0.25
- claude-opus-4-5Anthropic·对话·200K输入$5输出$25缓存读取$0.5
- claude-opus-4-6Anthropic·对话·1M输入$5输出$25缓存读取$0.5
- claude-opus-4-7Anthropic·对话·1M输入$5输出$25缓存读取$0.5
- claude-opus-4-8Anthropic·对话·1M输入$5输出$25缓存读取$0.5
- claude-opus-5Anthropic·对话·1M输入$5输出$25缓存读取$0.5
- claude-opus-5-5Anthropic·对话·1M输入$4输出$20缓存读取$0.2
- claude-sonnet-4-5Anthropic·对话·1M阶梯计价输入$3输出$15缓存读取$0.3
- claude-sonnet-4-6Anthropic·对话·1M输入$3输出$15缓存读取$0.3
- claude-sonnet-5Anthropic·对话·1M输入$2输出$10缓存读取$0.2
- claude-sonnet-5-5Anthropic·对话·1M输入$2输出$10缓存读取$0.2
- gemini-2.5-flashGoogle·对话·1.05M输入$0.3输出$2.5缓存读取$0.03
- gemini-2.5-flash-liteGoogle·对话·1.05M输入$0.1输出$0.4缓存读取$0.01
- gemini-2.5-proGoogle·对话·1.05M阶梯计价输入$1.25输出$10缓存读取$0.125
- gemini-3-flash-previewGoogle·对话·1.05M输入$0.5输出$3缓存读取$0.05
- gemini-3.1-flash-liteGoogle·对话·1.05M输入$0.25输出$1.5缓存读取$0.025
- gemini-3.1-pro-previewGoogle·对话·1.05M阶梯计价输入$2输出$12缓存读取$0.2
- gemini-3.5-flashGoogle·对话·1.05M输入$1.5输出$9缓存读取$0.15
- gemini-3.5-flash-liteGoogle·对话·1.05M输入$0.3输出$2.5缓存读取$0.03
- gemini-3.6-flashGoogle·对话·1.05M输入$0.75输出$3.75缓存读取$0.075
- gemini-3.7-flashGoogle·对话·1.05M输入$0.75输出$3.75缓存读取$0.075
- gemini-3.8-flashGoogle·对话·1.05M输入$0.75输出$3.75缓存读取$0.075
- gemini-omni-1.1-flashGoogle·对话·1.05M输入$1.5输出$9
- gemini-3-pro-imageGoogle·图像·66K输入$2输出$12
- gemini-3.1-flash-imageGoogle·图像·131K输入$0.5输出$3
- gemini-3.1-flash-lite-imageGoogle·图像·66K输入$0.25输出$1.5
- gemini-embedding-001Google·向量·2K输入$0.15输出—
- gemini-embedding-2Google·向量·8K输入$0.2输出—
- imagen-3.0-fast-generate-001Google·图像$0.02 / 次
- imagen-3.0-generate-001Google·图像$0.04 / 次
- nano-banana-pro-previewGoogle·图像·131K输入$2输出$12
- grok-4.20xAI·对话·1M阶梯计价输入$1.25输出$2.5缓存读取$0.2
- grok-4.3xAI·对话·1M阶梯计价输入$1.25输出$2.5缓存读取$0.2
- grok-4.5xAI·对话·500K阶梯计价输入$2输出$6缓存读取$0.3
- grok-4.6xAI·对话·500K阶梯计价输入$2输出$6缓存读取$0.5
- grok-4.7xAI·对话·500K阶梯计价输入$2输出$6缓存读取$0.5
- grok-code-fastxAI·对话·256K阶梯计价输入$1输出$2缓存读取$0.2
- deepseek-coderDeepSeek·对话·128K输入$0.14输出$0.28缓存读取$0.014
- deepseek-flashDeepSeek·对话·1M输入$0.3输出$1.2缓存读取$0.006
- deepseek-r1DeepSeek·对话·66K输入$0.55输出$2.19缓存读取$0.14
- deepseek-v3DeepSeek·对话·66K输入$0.27输出$1.1缓存读取$0.07
- deepseek-v3.2DeepSeek·对话·164K输入$0.28输出$0.4缓存读取$0.028
- deepseek-v4-flashDeepSeek·对话·1M输入$0.3输出$1.2缓存读取$0.006
- deepseek-v4-proDeepSeek·对话·1M输入$1.32输出$3.96缓存读取$0.044
- qwen-coderAlibaba·对话·1M输入$0.3输出$1.5
- qwen-maxAlibaba·对话·31K输入$1.6输出$6.4
- qwen-plusAlibaba·对话·129K输入$0.4输出$1.2
- qwen-turboAlibaba·对话·129K输入$0.05输出$0.2
- qwen3-next-80b-a3b-instructAlibaba·对话·262K输入$0.15输出$1.2
- qwen3-next-80b-a3b-thinkingAlibaba·对话·262K输入$0.15输出$1.2
- qwen3-vl-235b-a22b-instructAlibaba·对话·131K输入$0.4输出$1.6
- qwen3-vl-235b-a22b-thinkingAlibaba·对话·131K输入$0.4输出$4
- qwen3-vl-32b-instructAlibaba·对话·131K输入$0.16输出$0.64
- qwen3-vl-32b-thinkingAlibaba·对话·131K输入$0.16输出$2.87
- qwen3.7-maxAlibaba·对话·992K输入$2.5输出$7.5缓存读取$0.5
- qwen3.8-flashAlibaba·对话·992K输入$0.15输出$0.47缓存读取$0.016
- qwen3.8-maxAlibaba·对话·992K输入$2输出$6缓存读取$0.25
- qwen3.8-omni-flashAlibaba·对话·992K输入$0.15输出$0.47缓存读取$0.016
- qwq-plusAlibaba·对话·98K输入$0.8输出$2.4
- glm-4-32b-0414-128kZhipu AI·对话·128K输入$0.1输出$0.1
- glm-4.5Zhipu AI·对话·128K输入$0.6输出$2.2
- glm-4.5-airZhipu AI·对话·128K输入$0.2输出$1.1
- glm-4.5-airxZhipu AI·对话·128K输入$1.1输出$4.5
- glm-4.5-xZhipu AI·对话·128K输入$2.2输出$8.9
- glm-4.5vZhipu AI·对话·128K输入$0.6输出$1.8
- glm-4.6Zhipu AI·对话·200K输入$0.6输出$2.2缓存读取$0.11
- glm-4.7Zhipu AI·对话·200K输入$0.6输出$2.2缓存读取$0.11
- glm-5Zhipu AI·对话·200K输入$1输出$3.2缓存读取$0.2
- glm-5-codeZhipu AI·对话·200K输入$1.2输出$5缓存读取$0.3
- glm-5.1Zhipu AI·对话·200K输入$1.4输出$4.4缓存读取$0.26
- glm-5.2Zhipu AI·对话·1M输入$1.4输出$4.4缓存读取$0.26
- glm-5.3Zhipu AI·对话·1M输入$1.4输出$4.4缓存读取$0.26
- glm-5.3-flashZhipu AI·对话·1.05M输入$0.15输出$0.5缓存读取$0.03
- kimi-k2.5Moonshot AI·对话·262K输入$0.6输出$3缓存读取$0.1
- kimi-k2.6Moonshot AI·对话·262K输入$0.95输出$4缓存读取$0.16
- kimi-k2.7-codeMoonshot AI·对话·262K输入$0.95输出$4缓存读取$0.19
- kimi-k3Moonshot AI·对话·1.05M输入$3输出$15缓存读取$0.3
- MiniMax-M2MiniMax·对话·200K输入$0.3输出$1.2缓存读取$0.03
- MiniMax-M2.1MiniMax·对话·1M输入$0.3输出$1.2缓存读取$0.03
- MiniMax-M2.1-lightningMiniMax·对话·1M输入$0.3输出$2.4缓存读取$0.03
- MiniMax-M2.5MiniMax·对话·1M输入$0.3输出$1.2缓存读取$0.03
- MiniMax-M2.5-lightningMiniMax·对话·1M输入$0.3输出$2.4缓存读取$0.03
- MiniMax-M3MiniMax·对话·1M输入$0.3输出$1.2缓存读取$0.06
- doubao-seed-2-1-pro-260628ByteDance·对话·256K输入$0.863输出$4.31缓存读取$0.172
- doubao-seed-2-1-turbo-260628ByteDance·对话·256K输入$0.431输出$2.16缓存读取$0.0862
- codestral-2508Mistral AI·对话·128K输入$0.3输出$0.9缓存读取$0.03
- ministral-14b-2512Mistral AI·对话·262K输入$0.2输出$0.2缓存读取$0.02
- ministral-3-14b-2512Mistral AI·对话·262K输入$0.2输出$0.2缓存读取$0.02
- ministral-3-3b-2512Mistral AI·对话·131K输入$0.1输出$0.1缓存读取$0.01
- ministral-3-8b-2512Mistral AI·对话·262K输入$0.15输出$0.15缓存读取$0.015
- ministral-3b-2512Mistral AI·对话·131K输入$0.1输出$0.1缓存读取$0.01
- ministral-8b-2512Mistral AI·对话·262K输入$0.15输出$0.15缓存读取$0.015
- mistral-large-2512Mistral AI·对话·262K输入$0.5输出$1.5缓存读取$0.05
- mistral-large-3Mistral AI·对话·262K输入$0.5输出$1.5缓存读取$0.05
- mistral-mediumMistral AI·对话·262K输入$1.5输出$7.5缓存读取$0.15
- mistral-medium-3Mistral AI·对话·262K输入$1.5输出$7.5缓存读取$0.15
- mistral-medium-3.5Mistral AI·对话·262K输入$1.5输出$7.5缓存读取$0.15
- mistral-smallMistral AI·对话·32K输入$0.1输出$0.3缓存读取$0.01
- mistral-tinyMistral AI·对话·32K输入$0.25输出$0.25缓存读取$0.025
- open-mistral-nemoMistral AI·对话·128K输入$0.3输出$0.3缓存读取$0.03
- voxtral-small-2507Mistral AI·对话·33K输入$0.1输出$0.4缓存读取$0.01
- codestral-embedMistral AI·向量·8K输入$0.15输出—缓存读取$0.015
- mistral-embedMistral AI·向量·8K输入$0.1输出—
计费常见问题
在官方价格之外还有其他费用吗?+
没有。这里展示的就是按 token 计费的价格,没有平台费或订阅费。
如何付款?+
在控制台为账户充值,请求费用实时从余额中扣除。
失败的请求收费吗?+
不收费。返回错误的请求不计费。
有大客户价格吗?+
如有大量或长期的用量需求,请通过控制台联系我们,我们会为你定制方案。