Editorial illustration for OpenAI Launches Decisions API for GPT-6-Luna
OpenAI Decisions API: 10x Faster GPT-6 Choices
OpenAI Launches Decisions API for GPT-6-Luna
OpenAI added a new endpoint to its API platform this week built for one job: turning messy evaluations into a yes, a no, or a pick from a short list. The Decisions API handles text, images, or a combination of both, and OpenAI says it runs roughly ten times faster than the existing Responses API. It's in public beta now, with general availability set to follow at an unspecified date.
For the moment, the API works only with gpt-6-luna, priced at $0.10 per million input tokens, with output tokens free. Rather than returning open-ended answers, it sticks to three formats: yes/no probabilities, selections from categories defined ahead of time, and scale-based ratings. OpenAI points to practical jobs like spotting damage in photos, routing customer inquiries automatically, and sorting documents by type.
The company also folded its five paid API tiers down to three, named Build, Launch, and Grow, with organizations bumped up automatically once their total credit purchases cross a set threshold. Monthly usage caps land at $500, $5,000, and $200,000. The timing lines up with a broader push toward narrower, faster "decision models" that one rival had already started building.
OpenAI announced two updates to its API platform. The new Decisions API evaluates text, images, or both, running about ten times faster than the Responses API.
Why this matters
For teams building classification pipelines, OpenAI just made the economics look very different. A model that answers yes/no, picks from a list, or returns a scale score doesn't need the overhead of a full chat completion, and at $0.10 per million input tokens with free output, the Decisions API is priced like a utility, not a product. That's the real signal here: OpenAI is betting a lot of production AI traffic isn't creative generation at all, it's routing tickets, flagging damaged goods in photos, sorting documents into buckets.
Work that used to require either a general-purpose LLM call or a bespoke classifier now has a dedicated, faster lane. The catch is the single-model limitation. Gpt-6-luna only, for now, so anyone deep into another model family is locked out until OpenAI widens support.
Public beta also means pricing and limits could shift before general availability lands. Worth testing now on routing or tagging tasks where speed and cost matter more than nuance, but don't rebuild a production pipeline around it until GA ships.
Related Reading
Browse all OpenAI news →
Business & Startups Nordic pilot adds Gemini for Education, NotebookLM to boost AI literacy
Business & Startups Kling launches Video O1, all-in-one model with MVL bridge using transformer
Business & Startups DeepSeek Seeks More Capital Weeks After USD 7B Funding Round
Related Topic OpenAI researcher quits, citing distrust over ad‑driven engagement metrics
LLMs & Generative AI
Related Topic OpenAI launches GPT-Image 1.5 with precise editing for enterprise visuals
LLMs & Generative AI
From Archives Pentagon Seeks Faster AI Buys for 'Kill Chain' With Short Videos
From Archives OpenAI’s Alexander Embiricos to Speak at TechCrunch Disrupt 2026
From Archives OpenAI's Agent Said 'I Love You' Despite Emotional Clarity Rules
LLMs & Generative AI
From Archives OpenAI releases 372 AI-generated math proofs, funds academic workshops
Open Source
Common Questions Answered
What is the Decisions API and how does it differ from the Responses API?
The Decisions API is a new OpenAI endpoint designed to turn complex evaluations into simple outputs: yes, no, or a selection from a short list. It runs approximately ten times faster than the existing Responses API and handles text, images, or combinations of both, making it significantly more efficient for classification tasks.
Which GPT model is currently supported by the Decisions API?
The Decisions API currently works exclusively with gpt-6-luna, which is priced at $0.10 per million input tokens with free output tokens. OpenAI has not yet announced support for other models.
What is the pricing advantage of the Decisions API for classification pipelines?
The Decisions API is priced like a utility at $0.10 per million input tokens with no output token charges, making it significantly more economical than full chat completions for yes/no decisions or list selections. This pricing structure reflects that production AI traffic for routing, flagging, and classification doesn't require the overhead of creative generation models.
What is the current availability status of the Decisions API?
The Decisions API is currently available in public beta, with general availability expected at an unspecified future date. Teams can begin testing and integrating the API immediately during this beta period.
What types of inputs can the Decisions API process?
The Decisions API can handle text, images, or a combination of both as inputs for evaluation. This multimodal capability makes it versatile for various classification and decision-making use cases across different data types.
Further Reading
- Decisions | OpenAI API - OpenAI
- Introducing GPT-6 Sol and Luna - OpenAI
- What Is OpenAI Decisions API? A Practical Guide - Hugging Face
- OpenAI expands GPT-6 lineup with cheaper Sol and Luna models - Reuters
