OpenAI Decisions API公测:10倍速结构化决策
OpenAI Decisions API Hits Public Beta With 10x Faster Typed Answers
Agent开发者的新利器,原生支持概率输出的结构化决策接口能大幅简化工具调用逻辑,建议关注其延迟表现并评估接入可行性。
OpenAI has released the Decisions API in public beta. It turns text and images into typed answers your code can branch on. OpenAI team states the OpenAI Decisions API runs about 10x faster than the Responses API. It targets a common pattern: prompt an LLM, then parse its text into a label.
OpenAI 已发布 Decisions API 的公开测试版。它将文本和图像转化为类型化的答案,供代码进行分支判断。OpenAI 团队表示,OpenAI Decisions API 的运行速度比 Responses API 快约 10 倍。它针对一种常见模式:提示 LLM(大语言模型),然后将其文本解析为标签。
TL;DR
TL;DR(太长不看)
- Size: GPT-6 Luna parameter count is not disclosed. Its model card lists a 1,050,000-token context window.
- Runs on: OpenAI-hosted API only, via POST /v1/decisions. No open weights, no self-hosting.
- Performance: About 10x faster than the Responses API, per OpenAI.
- Best: $0.10 per 1M input tokens, with no output, cache-read or cache-write charges.
- Bottom line:
- Best: fast, typed decisions with probabilities.
- Worst: 1 model, beta status, no independent evals yet.
- 规模:GPT-6 Luna 的参数数量未披露。其模型卡片列出了 1,050,000 token 的上下文窗口。
- 运行环境:仅限 OpenAI 托管的 API,通过 POST /v1/decisions 调用。无开源权重,不支持自托管。
- 性能:据 OpenAI 称,速度比 Responses API 快约 10 倍。
- 最佳价格:每 100 万输入 token 0.10 美元,无输出、缓存读取或缓存写入费用。
- 总结:
- 最佳:快速、类型化且带有概率的判断结果。
- 最差:仅支持 1 个模型,处于测试阶段,尚无独立的评估数据。
What is the OpenAI Decisions API?
什么是 OpenAI Decisions API?
The Decisions API is an OpenAI endpoint that evaluates text, images or both and returns typed answers. It does not generate prose. A request has 3 fields: model, input and questions. The only supported model today is gpt-6-luna. OpenAI expects general availability in the coming weeks.
Decisions API 是 OpenAI 的一个端点,用于评估文本、图像或两者,并返回类型化的答案。它不生成散文式文本。请求包含 3 个字段:model(模型)、input(输入)和 questions(问题)。目前唯一支持的模型是 gpt-6-luna。OpenAI 预计将在未来几周内实现全面可用。
How does the Decisions API work?
Decisions API 如何工作?
Each request carries shared evidence plus a list of questions. Each question has a unique name, a type and instructions. The response returns an answers array keyed by those names.
每个请求都携带共享的证据以及一个问题列表。每个问题都有唯一的名称、类型和指令。响应返回一个以这些名称为键的 answers 数组。
What are the 3 question types?
有哪 3 种问题类型?
- predicate: checks a condition and returns a probability from 0 to 1. Example: does a product photo show a crack, tear or dent?
- choice: picks 1 value from options you supply. It also returns per-option probabilities and a confidence field.
- score: rates input against ordered levels, indexed from 0. The score is a probability-weighted average of level indices.
- predicate(谓词):检查条件并返回 0 到 1 之间的概率。例如:产品照片是否显示裂纹、撕裂或凹陷?
- choice(选择):从你提供的选项中选取 1 个值。它还返回每个选项的概率和一个 confidence(置信度)字段。
- score(评分):根据有序层级对输入进行评级,层级索引从 0 开始。分数是层级索引的概率加权平均值。
OpenAI’s severity example makes the math concrete. Level probabilities of 0.1, 0.7 and 0.2 yield a score of 1.1. That value sits between ‘Workaround available’ and ‘Fully blocked’.
OpenAI 的严重性示例使计算具体化。层级概率分别为 0.1、0.7 和 0.2 时,得分为 1.1。该值位于“有变通方案”和“完全阻止”之间。
When should you use Structured Outputs instead?
何时应使用 Structured Outputs(结构化输出)?
OpenAI draws a clear line. Use Decisions for probabilities, choices or scores. Use Structured Outputs to fill your own JSON schema or write explanations. Use function calling when a model must request a tool call with arguments.
OpenAI 划清了明确界限。如需概率、选择或评分,请使用 Decisions;如需填充你自己的 JSON 架构或撰写解释,请使用 Structured Outputs;当模型必须请求带参数的工具调用时,请使用函数调用(function calling)。
How fast is it, and what is the evidence?
它的速度有多快,证据是什么?
OpenAI’s document claims about 10x faster responses than the Responses API. DevDay coverage put a decision near 150 ms, versus about 1.6 seconds for regular Luna calls. OpenAI has not published accuracy or calibration data for the endpoint. The docs advise setting thresholds with labeled examples from your own application.
OpenAI 的文件声称,其响应速度比 Responses API 快约 10 倍。DevDay 的报道指出,一次决策耗时接近 150 毫秒,而常规 Luna 调用耗时约 1.6 秒。OpenAI 尚未公布该端点的准确性或校准数据。文档建议,使用来自你自己应用的分层示例来设置阈值。
How much does it cost, and where can you deploy it?
成本是多少,可以在哪里部署?
With gpt-6-luna, input costs $0.10 per 1M tokens. There are no output-token, cache-read or cache-write charges. Regional processing premiums and long-context multipliers still apply. Luna’s model card prices prompts above 272K tokens at 2x input rates.
使用 gpt-6-luna,输入成本为每 100 万 token 0.10 美元。无输出 token、缓存读取或缓存写入费用。区域处理溢价和长上下文乘数仍然适用。Luna 的模型卡显示,超过 272K token 的提示词按 2 倍输入费率计费。
- Endpoint: POST /v1/decisions on the OpenAI API, plus a Playground.
- SDK minimums: Python 3.26.0, JavaScript 7.30.0, Go 3.73.0, Ruby 0.101.0, Java 4.78.0.
- Compliance: Zero Data Retention and HIPAA for eligible customers.
- Residency: United States and Europe (EEA plus Switzerland).
- Voice: decisions can drive actions via client delegation with the Live API.
- 端点:OpenAI API 上的 POST /v1/decisions,以及一个 Playground(游乐场)。
- SDK 最低要求:Python 3.26.0、JavaScript 7.30.0、Go 3.73.0、Ruby 0.101.0、Java 4.78.0。
- 合规性:零数据保留 (ZDR) 及针对符合条件的客户的 HIPAA 合规。
- 数据驻留地:美国和欧洲(EEA 加上瑞士)。
- 语音:通过 Live API 的客户委托,决策可以驱动操作。
How does it compare with TypeSafe Jev?
它与 TypeSafe Jev 相比如何?
The closest rival is TypeSafe Jev, launched September 15, 2026. Jev is a System One model that returns typed values with calibrated probabilities. TypeSafe prices input at $0.042 per 1M tokens, with free output. That makes OpenAI’s base rate about 2.4x higher. TypeSafe reports 70 to 500 ms end-to-end, measured from the US West Coast. Jev supports up to 255 choices but remains in early access. OpenAI’s edge is image input, compliance options and open beta access.
最接近的竞争对手是 TypeSafe Jev,于 2026 年 9 月 15 日发布。Jev 是一个 System One 模型,返回带有校准概率的类型化值。TypeSafe 的输入价格为每 100 万 token 0.042 美元,输出免费。这使得 OpenAI 的基础费率高出约 2.4 倍。TypeSafe 报告端到端延迟在 70 到 500 毫秒之间,测量地点为美国西海岸。Jev 支持最多 255 个选项,但仍处于早期访问阶段。OpenAI 的优势在于图像输入、合规选项和公开测试版访问权限。
| Feature | OpenAI Decisions API | TypeSafe Jev 1.13 | GPT-6 Luna via Responses API |
|---|---|---|---|
| Release | Beta, Oct 6, 2026 | Early access, Sep 15, 2026 | GA, Sep 22, 2026 |
| Underlying model | gpt-6-luna | Jev (System One model) | gpt-6-luna |
| Parameters | Not disclosed | Not disclosed | Not disclosed |
| Input | Text, images (inline base64) | Text and structured state | Text, images |
| Output | Probability, choice or score | Typed values with probabilities | Generated text (JSON via Structured Outputs) |
| Max options per question | Not disclosed | Up to 255 | Not applicable |
| Speed (vendor-stated) | About 10x faster than Responses API | 70 to 500 ms end-to-end | Baseline |
| Input price per 1M tokens | $0.10 | $0.042 | $0.10 |
| Output price per 1M tokens | $0 | $0 | $0.50 |
| Compliance | ZDR, HIPAA (eligible); US and EU residency | Not disclosed | EU data residency available |
| 功能 | OpenAI Decisions API | TypeSafe Jev 1.13 | GPT-6 Luna via Responses API |
|---|---|---|---|
| 发布状态 | Beta,2026 年 10 月 6 日 | 早期访问,2026 年 9 月 15 日 | GA,2026 年 9 月 22 日 |
| 底层模型 | gpt-6-luna | Jev (System One 模型) | gpt-6-luna |
| 参数 | 未披露 | 未披露 | 未披露 |
| 输入 | 文本、图像(内联 base64) | 文本和结构化状态 | 文本、图像 |
| 输出 | 概率、选择或分数 | 带概率的类型化值 | 生成的文本(通过 Structured Outputs 的 JSON) |
| 每问题最大选项数 | 未披露 | 最多 255 | 不适用 |
| 速度(厂商声明) | 比 Responses API 快约 10 倍 | 端到端 70 到 500 毫秒 | 基线 |
| 每 100 万 token 的输入价格 | $0.10 | $0.042 | $0.10 |
| 每 100 万 token 的输出价格 | $0 | $0 | $0.50 |
| 合规性 | ZDR、HIPAA(符合条件);美国和欧盟数据驻留 | 未披露 | 提供欧盟数据驻留 |
Key Takeaways
关键要点
- OpenAI Decisions API returns probabilities, choices and scores, not prose.
- Input costs $0.10 per 1M tokens; output is free.
- OpenAI claims about 10x faster than the Responses API.
- TypeSafe Jev is cheaper at $0.042 per 1M input tokens.
- OpenAI Decisions API 返回概率、选择和分数,而非散文式文本。
- 输入成本为每 100 万 token 0.10 美元;输出免费。
- OpenAI 声称其速度比 Responses API 快约 10 倍。
- TypeSafe Jev 更便宜,输入 token 价格为每 100 万 0.042 美元。
Check out the Technical details here. All credit goes to the researcher of this project. Also, feel free to follow us on Twitter and don’t forget to join our 150k+ML SubReddit and Subscribe to our Newsletter. Wait! are you on telegram? now you can join us on telegram as well.
在此查看技术细节。所有功劳归于该项目的研究者。此外,欢迎在 Twitter 上关注我们,别忘了加入我们拥有 15 万+成员的 ML SubReddit,并订阅我们的新闻通讯。等等!你还在 Telegram 上吗?现在你也可以加入我们的 Telegram 群组了。
更进一步:量化金融体系
看懂新闻只是起点——沿量化金融路径,把它变成能交付的工程能力