Decision Model Tracker

All models · First seen 19 Oct 2026 · Catalog updated 16 Nov 2026

Zoom Decisions

Zoom · Released 19 Oct 2026

A decision API in Zoom AI Services for high-volume classification, routing, triage, and tagging. It evaluates text and images against developer-defined outcomes and returns typed answers with probability estimates. Intended to sit between speech recognition, enterprise context, and an application's actions.

Maker
Zoom
Release date
19 Oct 2026
First seen
19 Oct 2026
Access
Zoom Decisions APIs in Zoom AI Services; developers can sign in at Zoom.ai. Request and response format was described as OpenAI-compatible.
Pricing (as publicly stated)
As stated at launch on 19 Oct 2026: $0.042 per million input tokens.
Modalities
text, images
Question types
yes/no, choice, score
Latency (as stated)
Zoom said that in an 15 Oct 2026 test, Zoom Decisions posted a 117 ms median latency, lower than the OpenAI and Microsoft decision APIs it tested through OpenRouter, and noted that the network paths differed.
Official announcement
https://www.zoom.com/en/ai-research/introducing-zoom-decisions-apis-accurate-speech-reliable-action/

Benchmark claims

These are claims as published by the maker or in the linked official post. We have not independently rerun every benchmark.

In Zoom's internal evaluation, Zoom Decisions averaged 81.3% accuracy across five text benchmarks (Contested Fast Decisions, JevBench, RAGTruth, JudgeBench, and ContractNLI). Zoom said both of its models outperformed Microsoft-Decision-1, TypeSafe's Jev, and OpenAI's Decisions API on those tests, and asked readers to treat the figures as Zoom internal results, not end-to-end agent performance.

Source: https://www.zoom.com/en/ai-research/introducing-zoom-decisions-apis-accurate-speech-reliable-action/

Names and marks belong to their owners. This page is an independent summary, not a product page from Zoom.