OpenAI announces Decisions API, which returns predefined answers with confidence scores, similar to TypeSafe's Jev, in limited preview with no pricing details

Predefined answers with confidence scores are now available, simplifying your decision-making processes.

OpenAI has introduced a new Decisions API, currently in limited preview, that provides predefined answers along with confidence scores. This offering aims to simplify decision-making processes by delivering structured outputs, similar to TypeSafe's Jev product. The API is built on OpenAI's Luna infrastructure. Pricing details for the Decisions API have not yet been released. This new service allows developers to integrate deterministic responses into their applications, potentially reducing the complexity of managing varied AI outputs. The confidence scores are intended to help users gauge the reliability of the provided answers, enabling more informed integration into workflows.

The introduction of OpenAI's Decisions API suggests a market trend towards more predictable and reliable AI outputs, moving beyond the often open-ended responses of general-purpose language models. This focus on confidence scores indicates an effort to bridge the gap between AI capabilities and enterprise-grade applications where certainty and verifiability are paramount. Companies leveraging AI for critical functions may find this API a more suitable option for specific, rule-based decision automation.

This development directly challenges existing solutions like TypeSafe's Jev by offering a comparable feature set from a major AI provider. Developers and businesses considering how to implement AI for tasks requiring specific, validated outcomes will now have a direct alternative from OpenAI. The limited preview suggests OpenAI is gathering feedback to refine the offering, with potential implications for how AI is integrated into workflows requiring high levels of assurance and reduced variability.