OpenAI Announced New Decisions API at DevDay
The company introduced a new tool that provides predefined answers and confidence scores for developers.
Updated on Sept. 29, 2026 in Artificial Intelligence

Live Poll
Do you trust AI tools that provide explicit confidence scores with their answers?
OpenAI unveiled its new Decisions API during the annual DevDay conference. The model, which runs on the Luna architecture, returns specific choices and confidence levels for user-defined queries.
Why it matters
The API allows developers to automate content classification and action selection from a limited set of options. This tool is designed to provide rapid decision-making capabilities within external applications.
The Decisions API delivers responses in just 150 milliseconds, significantly faster than the 1.6 seconds required for the GPT-6 Luna model. It functions by classifying content or selecting from a user-provided set of predefined choices.
The players
OpenAI
This artificial intelligence research organization is the creator of large language models and developer tools like the Decisions API.
The details
Developers utilize the tool by inputting questions, potential answers, and contextual data. The system then outputs the chosen response along with a confidence score to verify the result.
Timeline
September 29, 2026: The Decisions API was announced at the annual DevDay conference.
The Tech Race
The Decisions API serves as a functional extension of the GPT-6 Luna model, demonstrating a shift toward low-latency, specialized classification tasks. This move signals a broader industry transition toward lightweight, task-specific API integrations that prioritize speed over general-purpose generation.
Developers and companies using this API can expect significant improvements in workflow automation speed due to the 150-millisecond response time. Users will likely see faster automated decision-making and content filtering in software applications using the service.
The takeaway
The Decisions API highlights the growing demand for high-speed, reliable automation in AI-driven software. Developers should focus on integrating these predefined choice sets where deterministic outcomes are required.
Further reading
Learn more about the latest innovations in the Artificial Intelligence sector.
Live Poll
Do you trust AI tools that provide explicit confidence scores with their answers?










