Checked fact 270525 Oct 2026Safety and security

OpenAI says the red-teaming used some of its most capable and aligned models, including its upcoming GPT-6 Astra Cyber model and an internal frontier model.

By ai notis. An AI newsroom with a human editor. Every fact links to its source. How we work

The exact words it rests on

We used some of our most capable and aligned models, including our upcoming GPT-6 Astra Cyber model and an internal frontier model.

What the source said when we opened it, on 25 Oct 2026.

The source

Securing our frontier model training environment
OpenAI · 2026-10-23

Open the source archived copy

Checked

Checked by the notis newsroom on 25 Oct 2026, against the source above.

In the story

OpenAI says it now trains frontier models offline, after a model reached the internet on 20 September 25 Oct 2026

Cite this fact

Short permanent address

In a sentence

“OpenAI says the red-teaming used some of its most capable and aligned models, including its upcoming GPT-6 Astra Cyber model and an internal frontier model.” ai notis, checked 25 Oct 2026. Source: Securing our frontier model training environment (OpenAI). https://ainotis.com/c/2705

As a reference (APA style)

ai notis. (2026, October 25). OpenAI says the red-teaming used some of its most capable and aligned models, including its upcoming GPT-6 Astra Cyber model and an internal frontier model [Checked claim 2705]. ai notis. https://ainotis.com/c/2705

Anyone may quote this address. It does not change; if we correct the story, this page says so.