OpenAI makes Decisions API generally available

OpenAI says the Decisions API helps applications and agents make near-real-time decisions from text, images and structured inputs, up to 10 times faster than GPT-6 Luna through the Responses API.[1] It returns a probability, choice or score that an application can use to decide what happens next.[1] OpenAI lists classification, request routing and deciding what action to take next as supported tasks.[1] Developers can pass native JSON inputs and include images in individual question instructions.[1] The API is available through a dedicated endpoint with GPT-6 Luna.[1]

Since the public beta, OpenAI added native integration with GPT-Live, which it says helps applications make fast, focused decisions during live conversations.[1] OpenAI reports that this integration reduced time to next action by more than 90%, from 1.4 seconds to 0.13 seconds.[1] The company also added prompt caching for reused context and inference residency in the United States and European Union.[1] It raised the maximum number of questions per request to 400, twice the previous limit, and says it has added inference capacity in several regions in Asia.[1]

OpenAI charges $0.10 per million input tokens and $0.01 per million cached input tokens; it says there are no output-token charges and billing is only for input tokens.[1] OpenAI says regional processing is available in the United States and European Union, and that Zero Data Retention is available to eligible customers.[1] These details give teams concrete cost and data-handling terms to check when evaluating a workload. The evidence does not specify the eligibility criteria for Zero Data Retention or provide a complete list of supported processing regions.[1]

OpenAI says the API delivered 40 times faster p95 performance than available frontier models, with comparable accuracy.[1] The company also reports that the API made AI-cost attribution four times faster and 30% cheaper than GPT-6 Luna, without sacrificing accuracy.[1] In another comparison, it says the API ran under 280 milliseconds p95, about three times faster and 90% lower cost per call than the previous approach, with no drop in customer outcomes.[1] OpenAI says its dots team uses the API to reduce response times and improve interactivity, while its Personal Finance team uses it to categorize transaction descriptions.[1]