OpenAI's Astra and Anthropic's Claude crack unsolved Enigma messages
Two cryptanalysts used frontier AI models to decode long-standing World War II ciphers that had resisted human efforts for decades.
What to know
- OpenAI's Astra and Anthropic's Claude Opus 5 each independently broke long-unsolved Enigma messages, demonstrating frontier LLMs' capability in specialized cryptanalysis and archival research.
- Astra's approach was autonomous—conducting its own research, finding context clues, and building a functional Enigma simulator—accomplishing in two days what would take human researchers weeks or months.
- Only seven unbroken Enigma messages remain from World War II, plus one where plaintext is known but the code unbroken; both recent breaks hint at AI's potential in historical cryptology and archival problem-solving.
Carter Leffen DeveloperJack Willis Cybersecurity executiveFrode Weierud Cryptanalyst and Crypto Cellar maintainer
How it unfolded 1 development · click the chart to see its coverage articlesposts
-
1
Weierud validates both codebreaks, notes seven messages remain unsolved
Crypto Cellar maintainer Frode Weierud confirmed both solutions and noted uncertainty about whether Astra accessed private archival collections not publicly hosted. He expressed awe at the results and reported that seven unbroken Enigma messages remain, plus one where plaintext is known but code unsolved.
“GPT–6 Astra is behaving like a very professional cryptanalyst and archive researcher.”
— Frode Weierud -
F
TechCrunch has just published an article about the AI-assisted breaking of the Enigma messages MVUEH and FMNGI.
-
-
background
Leffen cracks 2005 Enigma message with OpenAI's Astra — Developer Carter Leffen directed OpenAI's newest model Astra to search a database of Enigma messages for an unbroken one and decode it. Astra conducted its own archival research, found context clues, built an Enigma machine simulator, and recovered plaintext from a message that had baffled researchers for 21 years. Leffen also created an interactive website explaining the solution.
-
background
Willis breaks Enigma message using Claude Opus 5 — Cybersecurity executive Jack Willis used Anthropic's Claude Opus 5 model to decode an unsolved Enigma message, leveraging the known signature of a particular officer's name. He contacted Weierud to validate the result.