[OpenAI Developer Platform Adds Decisions API Testing Interface, Supporting Judgment, Selection, and Scoring] According to TestingCatalog, the OpenAI Developer Platform has added a Decisions API testing interface, allowing users to test three modes: Predicate, Choice, and Score. These modes are respectively used to estimate the probability of a condition being true, make selections from given options, and assign scores based on evaluation criteria. OpenAI's official documentation indicates that this API is currently in public testing phase, supporting both text and image inputs. It can be applied to content classification, request routing, and task prioritization. Currently, it only supports the gpt-6-luna model, with a base input cost of $0.10 per million tokens, and no fees for output tokens.