CYBERSECURITYTRACKER
TRACKING6,506 stories in this site build1,309 vulnerability news stories in this site build
Permanent story citation

GPT-6 Astra Scores 100% on ExploitBench as OpenAI Blocks PoC Exploit Requests

This page keeps the story as Cybersecurity Tracker first published it. If the tracker later corrects it, the correction appears below the original and never replaces it.

Back to newsStory 5677

As cited

Copy frozen at (site build).

vulnerabilities

GPT-6 Astra Scores 100% on ExploitBench as OpenAI Blocks PoC Exploit Requests

No summary had been written when this copy was frozen.

First seen by Cybersecurity Tracker

Source attribution

Correction

Correction recorded as of .

vulnerabilities

GPT-6 Astra Scores 100% on ExploitBench as OpenAI Blocks PoC Exploit Requests

OpenAI unveiled GPT-6 Astra, described as the world's most intelligent and aligned model, after the system reached a Critical cybersecurity capability threshold under the company's Preparedness Framework. The model scored 100% on ExploitBench, a benchmark measuring artificial intelligence (AI) security capabilities, while OpenAI restricted access to proof-of-concept exploit requests.

Why it matters: Security teams evaluating frontier AI systems need to understand OpenAI's capability assessments and access controls, as advanced models with high exploit-generation scores present novel attack surface considerations for organizations integrating them into operations.

Source published
First seen by Cybersecurity Tracker

Source attribution

Glossary