OpenAI made its Decisions API decision model widely available to customers
Deployment Provisional 70% confidence first seen
OpenAI released its Decisions API decision model more broadly to customers, aiming to support fast, low-cost option-selection workflows with latency targets of roughly 100–180 milliseconds. The update positions the API as a Jev competitor and allows developers to swap traditional LLM generation for decision-model inference, avoiding token-by-token decoding in applicable cases.