Astra and Opus just passed Turing’s other test

Two AI language models have successfully decoded previously unsolved messages that were encrypted using Enigma machines during World War II. A developer used OpenAI's Astra model to break a message that had remained uncracked since 2005 by conducting archival research and building a simulator of the Enigma machine, while another cryptanalyst used Anthropic's Claude Opus 5 model to break a different unsolved message. This is notable because Alan Turing, the computer pioneer famous for his test of artificial intelligence, actually spent his lifetime cracking Enigma codes for the United Kingdom during the war. According to a cryptology expert who maintains an archive of Enigma messages, there are only seven unbroken Enigma messages remaining, suggesting that AI models may soon solve all of them.