OpenAI's Decisions API Is Now Generally Available: Speed, Pricing and What's New
On October 22, 2026, OpenAI made the Decisions API generally available. Built on GPT-6 Luna, it answers bounded questions — a probability, a choice, or a score — up to 10x faster than GPT-6 Luna through the Responses API. Since its public beta, OpenAI added native GPT-Live integration, prompt caching, regional processing in the US and EU, and raised the per-request question limit to 400. It costs $0.10 per million input tokens, with cached input at $0.01 per million tokens and no charge for output tokens. [1]
SeekVero editorial · Published · Prepared with AI assistance; no hands-on testing claimed.
Prepared from cited sources. No hands-on testing claimed.
What the Decisions API does
Applications and agents make many small decisions throughout a workflow, such as how to classify an input, where to route a request, or which action to take next; individually simple, these decisions add up in latency and cost at scale. [1]
The Decisions API is built with GPT-6 Luna and focuses the model on questions with bounded answers, helping applications and agents make near real-time decisions from text, images, and structured inputs up to 10x faster than GPT-6 Luna through the Responses API. [1]
Developers define a question and its allowed answers, then provide context as text, images, or native JSON. The API returns one of three answer types: predicate (a probability estimate that a condition is true), choice (one of the supplied options, with probabilities and confidence), or score (a probability-weighted average of ordered rubric levels). [1]
What changed since the public beta
Since the Decisions API's public beta, OpenAI says it has added a native GPT-Live integration, prompt caching for reused context, and inference residency in the US and EU. [1]
Developers can now pass native JSON inputs, include images directly in question instructions, and ask up to 400 questions per request — twice the previous limit. OpenAI also added capacity in several regions in Asia to improve latency. [1]
Decisions now integrates natively with GPT-Live: a session can add a decisions trigger, for example on a transcript update, so a voice agent or avatar reacts to what a user just said without a separate request. [1]
Pricing and availability
Decisions API is generally available today through a dedicated Decisions endpoint built on GPT-6 Luna. [1]
| Option or question | What to know |
|---|---|
| Input tokens | $0.10 per million input tokens. [1] |
| Cached input tokens | $0.01 per million tokens for cached input. [1] |
| Output tokens | No charge — you only pay for input tokens. [1] |
| Regional processing | Available in the US and EU. [1] |
| Zero Data Retention | Available to eligible customers. [1] |
What early users reported
Box's VP of Product Management for Platform and Integrations said that with Box Extract and the Decisions API, it classifies business documents and routes them to the next step, and that in its evaluation the Decisions API delivered about 5x faster route selection on 30 certificate-of-insurance cases, with the same accuracy as the Responses API. [1]
A Datadog staff engineer said the Decisions API is helping score the risk of agent-generated commands before executing them, delivering about 40x faster p95 latency than available frontier models with comparable accuracy. [1]
A Ramp research engineer said the Decisions API made attributing AI costs to actual work about 4x faster and 30% cheaper than using GPT-6 Luna directly, without sacrificing accuracy. [1]
A Sierra engineer said it uses the Decisions API to choose the next action during live customer conversations, running at under 280ms p95 — about 3x faster and 90% lower cost per call than its previous approach, with no drop in customer outcomes. [1]
Speak's CTO said the Decisions API integration in GPT-Live reduced the time to next action by over 90%, from 1.4 seconds to 0.13 seconds, in its testing. [1]
Frequently asked questions
Is the Decisions API still in beta?
No. OpenAI made it generally available on October 22, 2026. [1]
How much does the Decisions API cost?
$0.10 per million input tokens, with cached input at $0.01 per million tokens; there is no charge for output tokens. [1]
How many questions can one request include?
Up to 400 questions per request, twice the previous limit. [1]
Where can data be processed?
Regional processing is available in the US and EU, and Zero Data Retention is available to eligible customers. [1]
What answer types does it support?
Three: predicate (a probability), choice (one of your supplied options), and score (a probability-weighted average of rubric levels). [1]
Next step
OpenAI's own docs give current pricing, caching rules and how to build a first decision.
Read OpenAI's Decisions API pricing and developer guideProduct demos and reference material
Sources
Cite this SeekVero guide
Keep the source credit and original evidence links with your summary. This example does not grant additional reuse rights.
Markdown
Source: [SeekVero — OpenAI's Decisions API Is Now Generally Available: Speed, Pricing and What's New](https://seekvero.com/en/guides/openai-decisions-api-ga)
HTML
Source: <a href="https://seekvero.com/en/guides/openai-decisions-api-ga">SeekVero — OpenAI's Decisions API Is Now Generally Available: Speed, Pricing and What's New</a>