🤖OpenAI Opens Fast Decision Model API to All Customers
TL;DR
OpenAI made its Decisions API generally available, cutting response time from 1-3 seconds to roughly 100-180 ms. It takes aim at TypeSafe's Jev, now the most-queried decision model on OpenRouter.
OpenAI made its Decisions API generally available, cutting response time from 1-3 seconds to roughly 100-180 ms. It takes aim at TypeSafe's Jev, now the most-queried decision model on OpenRouter.
Key Points
Built by retuning the existing ChatGPT Luna model for faster, cheaper inference rather than training a new one
Latency dropped from 1-3 seconds to about 100-180 ms; voice and visual modalities added
Early users include Box, Datadog, Ramp and Speak; Speak cut action gaps from 1.4s to 0.13s
Artificial Analysis Decision Index: Jev ranks 3rd, OpenAI's model 14th
TypeSafe is valued at about $7.6B and shipped Jev 1.14 last week
Why It Matters
Decision models are the glue for agent loops. OpenAI is trading benchmark rank for distribution, and the catch is it still trails Jev on independent indices.
Quick Facts
Comments
Be the first to comment
Enjoyed this article?
Get it daily. 7am. Free. Reads in 5 minutes.