Service update · OpenAI
OpenAI Decisions API reaches general availability, doubling the per-request limit to 400 questions
- Effective time
- Review updated
OpenAI’s GPT-6 Luna-powered Decisions API has moved from public beta to general availability. Its dedicated endpoint answers bounded questions using text, images or JSON, returning a probability, choice or score. Additions since beta include native GPT-Live integration, prompt caching, US and EU inference residency, native JSON inputs and images in question instructions. The limit has doubled to 400 questions per request. Voice apps can choose whether to clarify, route or escalate while a conversation is underway, and reused context can lower input costs. Input costs $0.10 per million tokens, or $0.01 for cached input, with no output-token charges. Cache writes have separate pricing; Zero Data Retention is available to eligible customers. Developers can start in Decisions Playground or follow the developer guide.