OpenAI deploys GPT-6 Astra at Critical cyber capability level, raises monitoring concerns

OpenAI has broadly deployed GPT-6 Astra, the first model to reach 'Critical level' cybersecurity capabilities, including the ability to find zero-day vulnerabilities. The company acknowledges that the model's advanced capabilities make it harder to monitor, raising dual-use concerns. This marks a significant escalation in AI-enabled offensive cyber capability, with implications for vulnerability discovery and defense.