Large Language Models

OpenAI Makes Decisions API Widely Available With Responses as Fast as 100ms

TechCrunchOct 24, 2026

OpenAI made its Decisions API broadly available on October 22, offering faster model-based choices and adding voice and visual inputs.

OpenAI made its Decisions API available to all customers on October 22, bringing its decision-focused model out of initial customer trials. The company introduced the API at its developer day on September 29. It is designed for tasks where a system must choose among options quickly, rather than generate a longer response.

OpenAI product lead Nikunj Handa told TechCrunch that the API uses ChatGPT Luna, reconfigured to reduce inference time and cost, rather than a model trained specifically for decision-making. Handa said response times that had taken one to three seconds were reduced to roughly 100, 120 or 180 milliseconds. The public version also adds voice and visual capabilities, while OpenAI says it has positioned inference resources geographically to reduce latency in different regions.

Early users included Box, Datadog, Ramp and Speak. Speak, an AI language-tutoring company, reported that using the API cut the time between actions from 1.4 seconds to 0.13 seconds. OpenAI has also used the system to improve responsiveness in its Dots voice assistant. Those examples show where the short-response format may be useful, but performance depends on the particular workflow.

The field is developing quickly, and published rankings do not put OpenAI at the top. In the AIM System One Index and Artificial Analysis Decision Index cited by TechCrunch, OpenAI’s model ranked eighth and fourteenth, respectively. TypeSafe’s Jev ranked seventh and third; Liquid’s D1 Ultra and Base Lab’s Elephant-1 led those rankings. TechCrunch also reported that TypeSafe’s models remained the most queried decision models on OpenRouter at the time of publication.

OpenAI’s API is therefore entering a competitive category where speed, cost, calibration and task-specific accuracy all matter. Handa said future work could include specialized training and further latency reductions, but did not give a release timeline for those changes.

Industry Impact

Faster decisions in interactive products

Enterprise software, voice assistants, language tutoring

Overview

For product managers and developers, the Decisions API offers a way to reduce delays in product steps that require choosing an action rather than generating a long response. Speak reported cutting the time between actions from 1.4 seconds to 0.13 seconds, though results depend on the workflow. The API’s voice and visual inputs broaden potential uses, while geographically positioned inference resources aim to reduce regional latency. Teams should weigh speed alongside cost, calibration, and task accuracy: cited rankings place OpenAI behind several competitors.

Key Impact Points

Broad availability moves the API beyond initial trials, giving customers a new option for latency-sensitive decisions.

Speak’s reported improvement illustrates potential user-experience gains, but does not establish results for other workflows.

Rankings show that speed is only one competitive factor; task accuracy, calibration, and cost also matter.

Opportunities

Build more responsive decision flows

Enterprise software, voice assistants, language tutoring

OpenAI’s public Decisions API creates a practical route for teams whose products must select an action quickly rather than produce long-form text. Developers can test it in interactive voice assistants, tutoring flows, and other decision-heavy product steps, using the reported latency gains as a reason to run a workflow-specific pilot—not as a guaranteed result. Added voice and visual inputs may support multimodal decision flows. Enterprise buyers should compare end-to-end latency, cost, calibration, and task accuracy against alternatives, since cited rankings place OpenAI behind several competitors. Commercial value would depend on measurable improvements in responsiveness and deployment economics; the article provides no pricing or quantified customer ROI.

Latency-sensitive product decisions
Problem: Delays between user actions can make interactive products feel unresponsive.
Solution: Pilot the Decisions API for product steps that require a quick choice, including voice or visual inputs where relevant.
Target Users: Developers and product teams building interactive assistants or tutoring experiences.
Business Value: Potentially shorter waits between actions; outcomes should be measured in the specific workflow.
Revenue Potential: Could support more responsive paid products or features, but the article provides no pricing or quantified return.

Reference link

https://techcrunch.com/2026/10/22/openai-makes-its-jev-competitor-widely-available-as-decision-model-competition-heats-up

Image Source Link

https://techcrunch.com/wp-content/uploads/2026/10/Screenshot-2026-10-22-at-12.17.11-PM-1.png?resize=1200,988

Community Discussion (0)

No discussions yet. Be the first to start the conversation!