Security Researcher Confirms OpenAI Silently Nerfed GPT-5.6 Cyber Capabilities

rez0__ · x · 2026-08-05

Security researcher @rez0 confirmed rumors regarding OpenAI's silent model adjustments. He clarified that the cyber capabilities of GPT-5.6 were not nerfed over the weekend, but were actually downgraded earlier today.

Original post →

More from Models

Models channel →