ResearchSeptember 22, 2026
2
š±
GPT-6 Astra Cracks 85-Year-Old Enigma
OpenAI's GPT-6 Astra model has decrypted an Enigma-encrypted message that had remained unsolved since 2005.
#OpenAI#GPT-6#cryptography#Enigma#autonomous agents

š„ What happened
OpenAI's GPT-6 Astra independently cracked the Enigma message MVUEH from July 10, 1941 ā a cipher that had resisted all human attempts since 2005. The AI achieved the break in just two days, after autonomously identifying archive sources at the German Bundesarchiv and building its own Enigma Bombe software.
š” Why it matters
The AI didn't just find the key ā it also detected transcription errors in the original and a rare rotor turnover at the 72nd letter. For comparison: a human cryptanalyst spent weeks on similar archive research ā GPT-6 Astra did everything autonomously in 48 hours.
ā” Our take
When an AI solves historical cryptography by doing archive research and building its own tools, that's not a benchmark win ā that's the moment agents start doing real research work. The question is no longer if, but when they take our jobs.
The title, summary and analysis of this item were produced automatically by an AI system and have not been editorially reviewed. They may contain errors, bias or omissions ā when in doubt, read the linked original source.
Deep Dives & Similar Intelligence

GPT-6 Cracks WWI Cipher From 1918
AI modelsSeptember 17, 2026
ki-daily.
OpenAI Slashes GPT-6 Prices by Half
AI modelsSeptember 22, 2026
ki-daily.
OpenAI Cracks 100 Math Problems
ResearchSeptember 22, 2026

OpenAI's $278B Burn Is the Wrong Headline
Business & TrendsSeptember 19, 2026

Claude Hacks OpenAI for $6,500 Bounty
Ethics & SecuritySeptember 18, 2026