OpenAI Cannot Rule Out Critical Cyber Capability in Astra
OpenAI published a security update saying its latest internal evaluations of Astra, one of its upcoming models, mean it cannot rule out critical cyber capabilities under its Preparedness Framework.
Transcript
OpenAI says it cannot rule out critical cyber capabilities in its upcoming model Astra.
On August seventh OpenAI said internal evaluations of Astra showed significant advancements in agent based coding and cybersecurity. Expert assessments followed and critical capability cannot be ruled out.
OpenAI's framework defines the Critical threshold as developing zero-day exploits in hardened systems without human intervention. Earlier models were assessed at the High threshold instead.
OpenAI is pausing internal activities involving Astra that do not meet strengthened security requirements. It will work with government agencies and safety organizations to test capabilities.
Inboxsmith helps small businesses handle calls and messages so nothing gets missed. Please like and subscribe for more news.
Sources
Every claim in this video comes from the top ranking coverage of this topic. The claims and where each one came from:
- Published: August 7, 2026(OpenAI's official announcement)
- Our latest internal evaluations of Astra, one of our upcoming models, over the past few days indicate significant advancements in agentic coding and cybersecurity.(OpenAI's official announcement)
- These results, in addition to expert assessments, have led us to conclude last night that we cannot rule out critical cyber capabilities under our Preparedness Framework.(OpenAI's official announcement)
- While we continue to benchmark and assess this model, our preliminary evaluations indicate strong enough performance that we cannot rule out Critical capability level at this time.(OpenAI's official announcement)
- Under our Preparedness Framework, a model reaches the Critical cybersecurity threshold if it can identify and develop functional zero-day exploits of all severity levels in many hardened real-world critical systems without human intervention, or can devise and execute end-to-end novel strategies for cyberattacks against hardened targets given only a high level desired goal.(OpenAI's official announcement)
- Previous models, including GPT-5.6-Sol, have been evaluated for frontier cyber capabilities and assessed at the High (rather than Critical) threshold.(OpenAI's official announcement)
- We are pausing internal activities involving Astra that do not yet meet these strengthened security control requirements.(OpenAI's official announcement)
- We will work with relevant government agencies and select AI safety organizations to test the capabilities for this model.(OpenAI's official announcement)