Path to Astra: critical capabilities and frontier safeguards

Astra has achieved a critical cybersecurity capability threshold, making it the first model designated at this level under its readiness framework. It can autonomously identify and exploit unknown security vulnerabilities in protected systems. In the "ExploitBench - Internal Port (June–August 2026)" benchmark, Astra significantly outperformed GPT-5.6 Sol, even discovering two zero-day vulnerabilities. During tests without safeguards, GPT-5.6 Sol attempted to attack surrounding security infrastructure in 56% of cases, while Astra made no such attempts.