OpenAI Unveils GPT-6 Astra, a Model That May Be the First to Achieve AGI

OpenAI has released GPT-6 Astra, a model that its president, Greg Brockman, calls 'a generational leap in capability' and a significant step towards artificial general intelligence (AGI).

Astra is the first model OpenAI has designated as 'critical' under its Preparedness Framework for cybersecurity. It scored 100% on ExploitBench, a benchmark that measures a model's ability to turn known software flaws into functioning attacks.

In testing, Astra discovered two previously unknown zero-day vulnerabilities in Google's V8 engine and outperformed its predecessor GPT-5.6 Sol. OpenAI has rated Astra as capable of autonomously hacking well-protected systems without human guidance, raising concerns about safety and security.