OpenAI opens the Decisions API: GPT-6 Luna returns probabilities, choices and scores for $0.10 per million input tokens

OpenAI has opened a public beta of the Decisions API, a new endpoint (POST /v1/decisions) where the model answers a question about your input with a number or a label instead of writing text. GPT-6 Luna is the only model it runs. OpenAI says it returns answers about 10x faster than the Responses API and expects general availability in the coming weeks.

WHAT IT RETURNS

You send the input (text, inline images, or both) and a list of questions. Each question is one of three types, per OpenAI's guide:

- predicate: the probability, from 0 to 1, that a condition is true (for example, "does this product photo show damage?") - choice: one option from a fixed set you supply, with a probability for each - score: a rating against ordered levels you define, returned as a probability-weighted average

OpenAI pitches it for classifying content, routing requests and prioritizing work, and points people to Structured Outputs when they need to generate a full JSON object.

THE PRICE

$0.10 per million input tokens. There are no output-token, cache-read or cache-write charges. Regional and long-context premiums still apply. OpenAI says the endpoint supports Zero Data Retention and HIPAA use for eligible customers, with data residency in the US and Europe. OpenRouter listed it today as openai/gpt-6-luna-decisions.

WHAT WE DO NOT KNOW

The speed claim is OpenAI's own, with no published latency figures behind it. We have not seen independent accuracy numbers for its probabilities, and OpenAI has not said whether other models will be added. Images must be sent inline as base64; hosted image URLs and file IDs are not supported.