OpenAI Decisions Playground Lets Developers Test Three Question Types

OpenAI’s Decisions Playground lets developers test how its public-beta API evaluates text and images before writing code, using three question types suited to classification, request routing and task prioritization.

A predicate question estimates the probability that a condition is true. A choice question selects from options the developer supplies. A score question rates the input against defined criteria. Developers can use the Playground to try different questions and inputs before putting them into an application.

The Decisions API currently supports only gpt-6-luna. Base input pricing is $0.10 per million tokens, and Decisions requests carry no output-token charge. Regional processing premiums and long-context pricing multipliers can raise the input cost.

OpenAI released the Decisions API in public beta on Oct. 6.