OpenAI Astra can spot zero-day flaws without a human.
OpenAI’s newest model has passed a “critical threshold” in cybersecurity, achieving perfection on an attack benchmark and finding two zero-day exploits.
OpenAI said that its latest AI model, called Astra, has achieved something that no other OpenAI model has done before. It has surpassed what the company calls a “critical threshold” in cybersecurity. Astra passed the ExploitBench test with flying colors — and it independently identified two previously unknown zero-day exploits, or vulnerabilities.
It did all of this without any human intervention. OpenAI said that it will make Astra publicly available “soon,” though some of its more powerful capabilities will be heavily regulated for high-risk users.
Source: techcrunch.com
Free course
Stop reading about AI — start building with it
The free Claude Code course: your first site, tool or game — no coding. No upsells, no cross-sells — nothing to buy here.
Start free →
Author
Evgenii Arsentev
PhD · Chief Executive Officer, digital health
Articles · Latest articles