4 September 2026
OpenAI releases GPT-6 Astra, rated Critical for cybersecurity risk
First reported
The Decoder ran this on , a day before the other 12 sources picked it up.
- OpenAI released GPT-6 Astra, its most capable model yet, scoring 100% on ExploitBench (a test of ability to find and develop security vulnerabilities) versus 78.5% for the previous GPT-5.6 Sol.
- The model found two previously unknown security flaws during testing and is restricted by default for enterprise users, with additional safeguards required for deployment.
Where they differ
Reported by The Decoder, SCMP, Computerworld, InfoWorld