Two previously undeciphered Enigma machine messages have been successfully decrypted using large language models. The breakthroughs were achieved by researchers utilizing AI tools developed by OpenAI and Anthropic to process historical archives.
Developer Carter Leffen employed OpenAI's Astra model to locate and decode an Enigma message that had remained unsolved since 2005. The model independently searched archives, developed a simulation of the Enigma machine, and reconstructed the original text within two days.
Separately, cybersecurity specialist Jack Willis used the Claude Opus 5 model from Anthropic to decrypt another long-standing mystery message. By providing specific contextual data, including an officer's name, Willis enabled the AI to successfully break the cipher.
Cryptologist Frode Weierud confirmed the accuracy of these findings. According to Weierud, there are currently only seven Enigma messages remaining that have not been deciphered, plus one message where the plaintext is known but the cipher remains unbroken.
