started · updated
OpenAI launches Decisions API powered by GPT-6 Luna
OpenAI has announced the full release of its Decisions API, a specialized interface built on the GPT-6 Luna model. Unlike traditional generative APIs designed for long-form text, the Decisions API focuses on structured, bounded answers to facilitate real-time decision-making. It allows applications to process text, images, and structured data to produce numerical probabilities, selections from predefined alternatives, or calculated scores.
The API is designed for high-frequency workflows such as automated document classification, request routing, priority determination, and software agent monitoring. OpenAI reports that this endpoint can deliver responses up to ten times faster than using the standard Responses API with GPT-6 Luna, significantly reducing latency for repetitive decision tasks.
Key technical updates include support for native JSON input, prompt caching for reused context, and the ability to submit up to 400 questions in a single request. OpenAI also noted the addition of GPT-Live integration and expanded service capacity in several Asian regions. Regarding data residency, the service now supports inference data residency in the United States and the European Union. Pricing is set at $0.10 per million input tokens, with cached input priced at $0.01 per million tokens; output tokens are free of charge.