Wes Roth · 2026-09-02 · notable
Wes Roth — 'GPT-6 Astra Just Went CRITICAL' on OpenAI's frontier safeguards
Wes Roth's September 2 episode covers OpenAI's 'Path to Astra' post, in which OpenAI says Astra is the first model to reach the Critical cyber level in its Preparedness Framework and restricts its advanced cyber features.
Wes Roth walks through OpenAI's decision to gate Astra's cyber features after the model hit the Critical threshold.
What is it?
'GPT-6 Astra Just Went CRITICAL...' went up on the Wes Roth channel on September 2, 2026. The subject is OpenAI's 'Path to Astra' post from the day before, which is already a separate entry in this feed. The 'GPT-6' framing in the title is Roth's own labelling, not a name OpenAI uses.
How does it work?
OpenAI's post is what the episode reacts to: Astra is the first model the company places at the Critical cyber capability level of its Preparedness Framework, based on autonomous zero-day discovery and exploit development on benchmarks including ExploitBench. OpenAI's response was to hold the advanced cyber features back at launch, opening them only to vetted testers and Daybreak Blue partners, and to publish a full system card.
Why does it matter?
A lab declaring its own model too capable at offensive security to ship openly is a first, and the argument over whether gated access is the right answer is happening now. Roth's channel covers frontier releases for a general audience, so the episode is a faster route into what the Critical designation means than reading the framework and the system card end to end.
Who is it for?
people following frontier model safety decisions
Try it
Watch at youtube.com/watch?v=qRNZMGc7TMc