Models
Securing our frontier model training environment
Published 2026-10-23 · Reported by OpenAI
OpenAI explains how we strengthened containment for frontier model training and evaluation, including offline environments and high-compute red-teaming.
<1 min recap · A full recap has not been written for this story yet. This is the publisher's own short description.
Source: OpenAI