OpenAI GPT–6 Astra breaks Enigma message that has resisted solution since 2005
Points and comments are a snapshot, not live.
OpenAI's GPT-6 Astra autonomously breaks a WWII Enigma message unsolved since 2005.
On 15 September 2026, Carter Leffer requested validation of a break of German Army Enigma message MVUEH from 10 July 1941. The message had resisted solution since 2005. GPT-6 Astra worked autonomously, developing Python and C++ Enigma simulator and Bombe software. It used a crib from the related message SIPVX and discovered the correct key (wheel order 253) and plaintext. The AI also located German Bundesarchiv file references RS 3-3/20a and RS 3-3/63b, which a human researcher had taken weeks to find.
What commenters are saying
Commenters split into two camps: those skeptical of the achievement's significance and those impressed. Skeptics argued the AI succeeded by brute force and dumb luck, requiring no skill or imagination, and that such PR-driven news will be used to boost upcoming IPOs. Others countered that many human inventions also arise from luck, and that this capability frees people from tedious work. Some noted the feat was impossible months ago and remains newsworthy, while a few questioned whether AI can design new encryption or only break known weak systems.