--- title: OpenAI's Decisions API is now generally available description: OpenAI's Decisions API is now generally available on its platform. It lets apps and agents classify text and images, evaluate inputs, and decide what happens... canonical: https://daily.dev/posts/openai-s-decisions-api-is-now-generally-available-ymb87tein twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: OpenAI's Decisions API is now generally available | daily.dev og:description: OpenAI's Decisions API is now generally available on its platform. It lets apps and agents classify text and images, evaluate inputs, and decide what happens... og:url: https://daily.dev/posts/openai-s-decisions-api-is-now-generally-available-ymb87tein og:image: https://api.daily.dev/og/posts/yMb87tEIN.png og:image:alt: OpenAI's Decisions API is now generally available og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # OpenAI's Decisions API is now generally available **[Collections](https://daily.dev/sources/collections)** · 2 min read · 3 upvotes · 0 comments ## Summary OpenAI's Decisions API is now generally available on its platform. It lets apps and agents classify text and images, evaluate inputs, and decide what happens next without generating a full model response. OpenAI says this makes apps and agents faster and cheaper to run. The API supports three decision types: Predicate, Choice, and Score. An example Score output reportedly put a 21% chance on a GPT-7 release this year, but that figure came from the model itself, not from OpenAI. ## Content OpenAI made its Decisions API generally available on October 22. It's a dedicated endpoint for bounded questions like routing, classification, and severity scoring, the kind of step where you need to decide what happens next and don't need a paragraph of generated text. Instead of free text, it returns a probability, a choice, or a score. Three decision types are available: predicate, choice, and score. It can classify text and images and evaluate inputs, which OpenAI says makes apps and agents faster and cheaper to run. ## Speed and pricing The API uses the gpt-6-luna model. OpenAI says it runs about 10x faster than the same model through the Responses API. - Input: $0.10 per million tokens - Cached input: $0.01 per million tokens - Output tokens: free ## How requests work A single request can hold up to 400 questions of mixed types, so something like ticket triage can happen in one call. A new `input_json` content part lets you pass structured data directly instead of stringifying it first. ## Early reactions FabianHedin said his team at Lovable has been trying the API and likes what they're seeing. @testingcatalog described it as a solution for building workflows around extremely quick decisions, and shared an example where the API estimated a 21% chance that the GPT-7 family ships this year. Their verdict on that number: "Looks right." The launch was also part of a busy day at OpenAI. @thsottiaux listed it alongside OpenResponses joining the Linux Foundation, some performance improvements, and Uber going all-in on Codex. ## Questions this post answers ### What is OpenAI's Decisions API used for? It lets an app or agent make near real-time decisions inside a workflow, such as classifying text and images or evaluating inputs, without generating a full model response. This is meant to make apps and agents faster and cheaper to run. It supports three decision types: Predicate, Choice, and Score. _Developers wiring decision logic into agents can track new API releases like this one on daily.dev._ ### What are the decision types supported by OpenAI's Decisions API? Three types are supported: Predicate, Choice, and Score. Predicate-style decisions likely return a yes/no style evaluation, Choice selects among options, and Score returns a numeric likelihood, as shown in an example output scoring the chance of a GPT-7 release. _Teams choosing how to structure agent workflows can follow API updates like this on daily.dev._ ## Community take How the wider developer community reacted, aggregated from 1 discussion and 3 comments across x (as of 2026-10-26). **TL;DR:** The small discussion found barely touches the Decisions API itself, instead drifting into a tangent about trusting automated agent-monitoring and whether guardrails hold up against adversarial red-teaming. **Sentiment:** 10% positive · 30% mixed · 60% skeptical **The pushback** - Some doubt whether automated guardrails/monitoring can be trusted without human oversight on real outputs. - Concern that red-teaming could bypass agent monitoring, undermining confidence in diagnostic results. **By community** - x (skeptical): Replies express doubt about trusting automated agent oversight and worry that monitoring can be circumvented by adversarial actors. **Open questions** - Can automated guardrails be trusted without human review of real outputs? - How well does agent monitoring hold up against deliberate attempts to bypass it? **Highlights** > @AnotherCodingX Red team getting around the agent monitoring is the big one for me. Diagnostic wins don't mean much if you can't see what the agent does once someone tries to own it > — [harleyfoote\_ on x](https://x.com/harleyfoote_/status/2113611904652390434) > @AnotherCodingX Watching agents monitor other agents feels like a game of cat and mouse. Do you trust automated guardrails right now, or do you still want eyes on real outputs? > — [BreakPassAI on x · 1 comments](https://x.com/BreakPassAI/status/2113615142755639393) > @BreakPassAI It’s a great question. I currently adhere to the structure of agents monitoring the outputs of my evals. I think it’s not a terribly far departure from a traditional critic pattern > — [AnotherCodingX on x](https://x.com/AnotherCodingX/status/2113623969333219532) **Source threads** - [x](https://x.com/AnotherCodingX/status/2113591289606553761) · 2 points · 3 comments ## Similar posts on daily.dev - [OpenAI Pricing in 2026 for Individuals, Orgs & Developers](https://daily.dev/posts/openai-pricing-in-2026-for-individuals-orgs-developers-aeizgj8sw) · finout · 1 upvotes · 0 comments --- Tags: [#ai-agents](https://daily.dev/tags/ai-agents), [#architecture](https://daily.dev/tags/architecture), [#openai](https://daily.dev/tags/openai) [View this post on daily.dev](https://daily.dev/posts/openai-s-decisions-api-is-now-generally-available-ymb87tein) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"OpenAI's Decisions API is now generally available","url":"https://daily.dev/posts/openai-s-decisions-api-is-now-generally-available-ymb87tein","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/openai-s-decisions-api-is-now-generally-available-ymb87tein"},"datePublished":"2026-10-22T22:33:39.441Z","dateModified":"2026-10-26T15:27:14.788Z","description":"OpenAI's Decisions API is now generally available on its platform. It lets apps and agents classify text and images, evaluate inputs, and decide what happens...","image":"https://pbs.twimg.com/media/HVRJxZfX0AAj_-v.jpg","thumbnailUrl":"https://pbs.twimg.com/media/HVRJxZfX0AAj_-v.jpg","isAccessibleForFree":true,"articleSection":"Collections","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Collections","logo":"https://media.daily.dev/image/upload/s--fk_6ycEi--/f_auto,q_auto/v1780996001/logos/collections?_a=BAMAMiWQ0","url":"https://daily.dev/sources/collections"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/openai-s-decisions-api-is-now-generally-available-ymb87tein","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":3},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"ai-agents,architecture,openai","timeRequired":"PT2M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Collections","item":"https://daily.dev/sources/collections"},{"@type":"ListItem","position":3,"name":"OpenAI's Decisions API is now generally available"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/openai-s-decisions-api-is-now-generally-available-ymb87tein#faq","mainEntity":[{"@type":"Question","name":"What is OpenAI's Decisions API used for?","acceptedAnswer":{"@type":"Answer","text":"It lets an app or agent make near real-time decisions inside a workflow, such as classifying text and images or evaluating inputs, without generating a full model response. This is meant to make apps and agents faster and cheaper to run. It supports three decision types: Predicate, Choice, and Score. Developers wiring decision logic into agents can track new API releases like this one on daily.dev."}},{"@type":"Question","name":"What are the decision types supported by OpenAI's Decisions API?","acceptedAnswer":{"@type":"Answer","text":"Three types are supported: Predicate, Choice, and Score. Predicate-style decisions likely return a yes/no style evaluation, Choice selects among options, and Score returns a numeric likelihood, as shown in an example output scoring the chance of a GPT-7 release. Teams choosing how to structure agent workflows can follow API updates like this on daily.dev."}}]} ```