OpenAI says its unreleased Astra model may cross the 'Critical' cyber threshold — and pauses internal work on it
On 2026-08-07 OpenAI disclosed that internal evaluations of Astra, an upcoming model, show agentic coding and cybersecurity performance strong enough that it can no longer rule out the Critical cybersecurity capability level in its Preparedness Framework — a first; every prior frontier model including GPT-5.6-Sol topped out at High. Critical is defined as autonomously finding and weaponizing zero-days in hardened real-world systems, or executing novel end-to-end attack campaigns from only a high-level goal. OpenAI responded with isolated testing environments, restricted network and tool access, stronger weight encryption, sandboxed execution, universal monitoring for risky actions across all agentic applications, a pause on some internal Astra work, and plans to bring in government agencies and external safety orgs.
Source
↳ Follow the thread