The general-availability release is available with gpt-6-luna, according to the company’s API changelog.
The API now accepts up to 400 questions in one request, up from 200 during its public beta. OpenAI also added prompt-cache discounts for shared input across requests, support for images in individual question instructions, and direct structured-data input through input_json.
The changelog presents these as changes since the public beta and links to a separate list of updates. It does not state Decisions API pricing, rate limits, regional availability, eligibility requirements, measured latency, or comparative quality results. It also does not establish how the new 400-question limit performs for specific workloads or whether all capabilities have identical availability across OpenAI accounts.