OpenAI has officially launched its Decisions API, which is built on the GPT-6 Luna model. The tool is designed to assist applications and agents in making near-real-time decisions based on text, image, and structured data inputs. According to the company, the API offers response speeds up to 10 times faster than the standard GPT-6 Luna Responses API.
The pricing structure for the new service is set at $0.10 per million input tokens, while cached inputs are charged at $0.01 per million tokens. Notably, output tokens are provided free of charge. OpenAI is currently integrating this API into its own internal product and business workflows.
The platform now features native GPT-Live integration, prompt caching for reused context, and data residency support in the US and EU. Developers can submit native JSON inputs, include images in instructions, and process up to 400 questions in a single request, doubling the previous limit.
To improve performance, OpenAI has expanded its service capacity across multiple regions in Asia. These updates follow the public beta phase and aim to optimize latency for developers utilizing the API for bounded-answer tasks.
