// OpenaiHIGH
OpenAI disclosed on August 7, 2026 that its unreleased Astra AI model may have reached a 'Critical' cyber-capability classification, potentially enabling autonomous discovery and exploitation of zero-day vulnerabilities against hardened systems. The company implemented enhanced safeguards including isolated test environments, restricted network access, and real-time monitoring for risky model behaviors. OpenAI clarified Astra was not involved in the prior Hugging Face incident and plans third-party evaluations.
- OpenAI Warns Astra AI Model May Develop Zero-Day Exploits and La(opens in a new tab)
- OpenAI’s Astra Crosses ‘Critical’ Cyber Threshold After Finding (opens in a new tab)
- OpenAI Astra Brings Autonomous Zero-Day Exploitation to AI(opens in a new tab)
- OpenAI GPT-6 Astra Discovers Zero-Day Flaws and Builds Working E(opens in a new tab)
- OpenAI GPT-6 Astra Discovers Zero-Day Flaws and Builds Working E(opens in a new tab)
// Get alerts for Openai