At its 2026 Developer Day, OpenAI announced the Decisions API, designed for real-time, low-latency classification and routing, returning results in about 150 milliseconds — roughly 10x faster than regular API calls to Luna. Built on the small model Luna, it focuses on user-defined questions with a finite set of answers, returning structured decisions for uses such as customer service ticketing, content moderation, agent orchestration decision nodes, and real-time business rule assistance.