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 these are simple, but at scale their latency and cost add up. Today, OpenAI is making the Decisions API generally available. Built with GPT‑6 Luna, it focuses Luna's intelligence 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.
Since the public beta, OpenAI has added a native GPT‑Live integration, prompt caching for reused context, and inference residency in the US and EU. Developers can pass native JSON inputs, include images in question instructions, and ask up to 400 questions per request, twice the previous limit. Latency has also improved, with additional capacity in several regions in Asia.
Early adoption has shown broad applicability. Across OpenAI, teams use the API to reduce response times and improve interactivity, categorize transaction descriptions, and classify coverage by theme and whether it reinforces or challenges key narratives. Customers use it for everything from attributing AI costs accurately to making language-learning experiences more interactive.
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, a probability, a choice, or a score, that applications use to determine what happens next. With GPT‑Live, the Decisions API helps voice applications make fast, focused decisions during live conversations, such as whether a support agent should ask a clarifying question, route a request, or escalate an issue.
Source: OpenAI Blog · Summarized by HeadlinesBriefing