AI Models Crack Long-Unsolved Enigma Messages
New AI models, Astra and Opus, have successfully decoded Enigma-encrypted messages that Alan Turing's team could not solve during World War II. The achievement was announced on September 25, 2026.
Advanced artificial intelligence models, Astra and Opus, have successfully deciphered historical Enigma-coded messages that remained unbroken by Alan Turing's wartime cryptanalysis team. The breakthrough was reported on September 25, 2026.
According to technology analysts, Astra, developed by OpenAI, and Opus, from Anthropic, utilized sophisticated AI algorithms and archival research to uncover encrypted communications that had baffled codebreakers for decades. These messages were typically unsolvable due to transcription errors or mistakes made by the original encoders.
Alan Turing is widely recognized for the Turing Test, designed to assess machine intelligence. However, his critical wartime contribution involved breaking the German Enigma code at Bletchley Park. While Turing and his colleagues built the Bombe machine, an early computing device, to decrypt these messages, a small number proved intractable.
By employing AI models capable of independent research and contextual analysis, these systems have now effectively completed Turing's codebreaking efforts. This success unlocks previously inaccessible historical communications, offering new insights into World War II intelligence.