DigestAI news desk

Cut through the AI noise.

Generative AI & Modelsupdated 3 min read

OpenAI opens Decisions API to all customers, adding voice and visual modalities

OpenAI made its Decisions API model generally available on Thursday, after announcing it at developer day on September 29. The API offers a fast, cheap decision‑making model that can classify options and return confidence scores. It is built by reconfiguring ChatGPT Luna, shrinking inference latency from “one to three seconds” to “100, 120, 180 milliseconds” and adding voice and visual modalities.

2 sources

What you can do with it

AI at Work →
OpenAI Decisions API by OpenAI

Get instant, typed decisions in apps with OpenAI Decisions API

Returns probability, choice or rubric scores for up.

Who for
marketers
Cost
paid from $0.10 per million input tokens
Effort
minutes

Use it for

  • Classify content
  • Route requests
  • Prioritize work

Watch out Check the details on the maker's page first

OpenAI Decisions API in the tool directory

Key points from the news

  • OpenAI released the Decisions API to all customers on Thursday, after a September 29 developer‑day announcement.
  • The API, built on reconfigured ChatGPT Luna, runs inference in 100‑180 ms (down from one‑to‑three seconds) and adds voice and image support.
  • Box, Datadog, Ramp and Speak report latency cuts, with Speak dropping action time from 1.4 s to .13 s.

Early trials with Box, Datadog, Ramp and Speak show the API can replace standard LLM calls. Speak’s CTO Andrew Hsu said the latency dropped from 1.4 seconds to .13 seconds. OpenAI also used the model to speed up its Dots voice assistant. TypeSafe, the startup behind the Jev model that sparked the decision‑model category, is valued at $7.6 billion and still leads OpenRouter queries. Benchmarks place OpenAI’s Decision Model eighth on the AIM System One Index and fourteenth on the Artificial Analysis Decision Index, while its latest Jev model tops the JevMan PacMan test. Handa envisions future speeds of around 60 milliseconds for per‑frame video analysis.

Model pages: Jev → · d1-ultra →

Full story from TechCrunch AI · by Tim FernholzOpen source ↗

OpenAI makes its Jev competitor widely available as decision model competition heats up

TechCrunch AI · 22 October 2026

Loading the full article…

This text was published by TechCrunch AI and written by Tim Fernholz. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗

Coverage and discussion

2sources
Topics · follow one to build your own front page
OpenAITypeSafeBoxDatadogRampSpeakDecisions APIJevD1 UltraElephant-1LunaNikunj HandaAndrew HsuDiogo AlmeidaFelix Wunderlich

The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.

Comments

via GitHub Discussions

More in Generative AI & Models

All →

Related stories