OpenAI Decisions API
OpenAI's Decisions API: GPT-6 Luna answers Noul, Choice and Score questions about text, JSON or images with a probability per option.
This summary is based on how curated lists describe the project; its README was not available to review.
Overview
The OpenAI Decisions API was announced at DevDay 2026 and has been in public beta since October 2026, according to the list summary. GPT-6 Luna answers Noul, Choice and Score questions about text, JSON or image state, returning a probability for every option and no generated text, the same shape popularised by Jev. It is also available on OpenRouter as GPT-6 Luna Decisions.
No further documentation was fetched, so pricing, limits, latency and calibration are unknown here. Its significance is that a major model provider now offers the decision primitive directly, including image inputs, which hosted Jev did not support at the time.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Yes / no | Yes/no question about the state | yes · no |
| Choice | Pick one of the supplied options | — |
| Score | Place the state on ordered levels | — |
How it works
- Send text, JSON or image state with typed questions.
- GPT-6 Luna returns per-option probabilities.
- Branch in application code on the probabilities.
Strengths
- Supports image state as well as text and JSON.
- Available directly and via OpenRouter.
Caveats
- Public beta; details beyond the summary unknown.
- Hosted, proprietary model.
At a glance
- Type
- Service
- Status
- beta
- Runtime
- hosted API
- Has code
- no
Use cases and integrations
How curated lists describe it
Announced at DevDay 2026 and in public beta since October 2026: GPT-6 Luna answers Noul, Choice, and Score questions about text, JSON, or image state with a probability for every option and no generated text. Also on [OpenRouter](https://openrouter.ai/openai/gpt-6-luna-decisions) as GPT-6 Luna Decisions.
AnotiaWang/awesome-decision-models
Frequently asked questions
Which model powers it?
GPT-6 Luna, per the list summary.
Does it accept images?
Yes, text, JSON or image state.
Is it on OpenRouter?
Yes, as GPT-6 Luna Decisions.
Related projects
Simon Willison's LLM plugin that adds Jev as a model for yes/no, choice and score questions from the command line.
PluginproductionPythonGo integration bringing TypeSafe Choice, Score and Noul to Google's adk-go agents, with a typed client, routing agent and callbacks.
ProjectbetaGoRuby gem for probabilistic conditionals like Feelings(msg).like?("urgent"), answered by Jev or other decision models.
ProjectbetaRubyRust CLI and library that asks Jev typed questions via OpenRouter and combines the probabilities with fuzzy-logic rules into decisions.
CLIbetaRustElixir DSL (Gut.feel) that picks an Elixir value for a subject and question, with a Jev evaluation adapter among LLM options.
ProjectbetaElixirIndependent OpenClaw plugin exposing one explicitly invoked typesafe_decide tool that sends supplied state and typed questions to Jev.
HookbetaTypeScript