OpenAI Decisions API goes live: fast AI verdicts with no output charge
By Anika SharmaAI & Innovations Editor, Peepals Tech
Representative image: Pexels / Pixabay
OpenAI made its Decisions API generally available on October 22, 2026, returning typed AI answers from GPT-6 Luna at $0.10 per million input tokens.
What Happened
OpenAI made its Decisions API generally available to developers on October 22, 2026, taking it out of a public test phase that began on October 6. The service lets software ask an AI model short, structured questions, such as "is this message spam?" or "which of these four categories fits?", and get back a typed answer instead of a written reply. It runs on OpenAI's GPT-6 Luna model, and the company says it can return answers up to ten times faster than asking the same model through its standard interface.
Key Facts
- Generally available: October 22, 2026, after a public beta that opened on October 6.
- Model: GPT-6 Luna is the only supported model.
- Price: $0.10 per million uncached input tokens and $0.01 per million cached input tokens; output is not billed, according to OpenAI.
- Answer types: a probability that something is true, a choice among options, or a score against a set of criteria.
- Capacity: up to 400 questions in one request, up from 200 during the beta.
- Added since the beta: image inputs, prompt caching, integration with GPT-Live, and the option to process requests in the US or the EU.
- Early users named by OpenAI include Box, Datadog, Ramp, Sierra and Speak.
Why It Matters
An API, or application programming interface, is the connection that lets one piece of software use another. Most AI APIs today return text: a developer sends a prompt and the model writes an answer. That is slow and costly when a program only needs a quick verdict, such as whether to flag a transaction, route a support ticket or approve a piece of content. The developer must then parse the written answer to pull out the decision.
A "decision model" skips the writing. It returns a structured result directly, such as a probability between 0 and 1 or the name of one option from a list. Because the model does not generate a long reply, the answer comes back faster and, under OpenAI's pricing, the developer pays only for the input. A token is a small piece of text, roughly a word or part of a word, and AI services usually charge per token.
This matters most for AI "agents", programs that carry out multi-step tasks on their own. An agent may need to make dozens of small choices along the way, and each slow, costly call adds up. Cheaper and faster decisions make such systems more practical for businesses.
The launch also shows a new competitive front. A start-up called TypeSafe AI built a decision model, Jev, and OpenAI's move puts the largest AI company directly into that space. For developers, more competition can mean lower prices; for smaller firms, it can mean a powerful new rival.
For India, where many software companies build customer service, fintech and content-moderation tools for clients worldwide, cheaper classification at scale could lower the cost of running AI features. Indian developers should still weigh where data is processed: OpenAI offers processing in the US or the EU, which matters for firms handling personal data under India's data protection law.
| Feature | Detail |
|---|---|
| Status | Generally available from October 22, 2026 |
| Model | GPT-6 Luna |
| Input price | $0.10 per million tokens (uncached) |
| Output price | Not billed |
| Questions per request | Up to 400 |
Impact
Short-term: Developers can now use the Decisions API in production apps. Companies already testing it during the beta can move to wider use.
Long-term: If structured decision calls become standard, AI features such as spam filters, fraud checks and ticket routing could become cheaper and faster, and rival AI companies are likely to offer similar services.
Who is affected: software developers and companies building AI agents and automation, including in India's IT and start-up sector; decision-model start-ups such as TypeSafe AI; and businesses that use AI for classification and moderation.
Key Takeaway
OpenAI's Decisions API, generally available from October 22, 2026, returns typed answers from GPT-6 Luna at $0.10 per million input tokens with no output charge, aiming to make quick AI decisions faster and cheaper for apps and agents.
Questions and Answers
What is the OpenAI Decisions API?
It is a service, generally available since October 22, 2026, that lets software ask OpenAI's GPT-6 Luna model short structured questions and get back a probability, a choice or a score instead of a written answer.
How much does the Decisions API cost?
OpenAI charges $0.10 per million uncached input tokens and $0.01 per million cached input tokens. Output tokens are not billed.
How is it different from a normal chatbot API?
A normal API returns generated text that a program must interpret. The Decisions API returns a typed result directly, which OpenAI says can be up to ten times faster for these tasks.
Who is using it?
OpenAI names Box, Datadog, Ramp, Sierra and Speak among its users.
Disclaimer: Prepared by the Peepals newsroom from publicly available sources with AI assistance. Information is accurate to the best of our knowledge at the time of publication and may change. Images may be representative. Not professional advice. Report an error via our contact page.
By Anika Sharma, AI & Innovations EditorCovers AI & Innovation and more for Peepals Tech. All stories by Anika Sharma
Moonshot AI releases Kimi K3.1 Pro, a 2.8-trillion-parameter open-weight model
American Express to cut 5% of staff to prepare for AI
Anika Sharma is a pen name of the Peepals Global Editorial Team, which researches, fact-checks and publishes every story. The portrait is an illustration.
