GPT-6 Astra cracks 83-year-old Enigma cipher after marathon autonomous session
A Bloomberg developer used OpenAI's GPT-6 Astra to decrypt a 1941 Wehrmacht Enigma message that had remained unsolved for over eight decades. The AI worked for about ten hours, combining historical archive searches, cryptanalysis code, and an Enigma simulator to find the plaintext. The result has not yet been independently verified, and the breakthrough relied on a known town name and the Enigma machine's property of never encrypting a letter as itself.
The decrypted message originated from a German soldier on July 10, 1941, who reported his location in Rosenow and requested marching orders. The 82-character transmission had remained unsolved for 83 years, listed in an archive of intercepted German army messages. Carter Leffen, a Bloomberg product development coach, ran the "GPT-6 Astra Extra High" variant for roughly ten hours, during which the AI handled archive searches, cryptanalysis code, an Enigma simulator, and parallel key testing.
The breakthrough leveraged two factors: a previously decrypted message from the same day containing "Rosenow" twice, and the Enigma machine's known weakness of never encrypting a letter as itself. The recovered plaintext included phrases like "Sofort Funkantwort" and "Angabe des Marschweges," along with operator typos such as "BTTE" for "BITTE." All code and verification tools are publicly downloadable, though independent verification remains pending.
This achievement could signal a shift in how historical cryptanalysis is conducted, potentially opening previously unsolvable archives to fresh scrutiny. If verified, it may demonstrate that AI systems can meaningfully assist human experts in complex, multi-step research tasks. However, the lack of independent verification means the result could also highlight the need for rigorous review standards as AI-driven discoveries become more common in specialized fields.