A breakthrough in historical cryptography: Anthropic's new language model, Claude Fable 5.1, has successfully deciphered the famous Cyphral Distich cryptogram by Scottish writer Thomas Urquhart. This puzzle, consisting of two lines of 32 numbers each, had remained unsolved for over 370 years, since its publication in the treatise Logopandecteision in 1653. I learned of this achievement from the results of the independent Vals AI project, which was testing the model's capabilities.
For decades, the cipher baffled the best experts. As early as 1899, it was brought up for public discussion, and in the 20th century it firmly entered cryptographic literature. Researcher Klaus Schmeh even placed it 28th on his list of the 50 most challenging undeciphered messages in history. All previous attempts were based on frequency analysis and substitution ciphers, but none yielded results.
The key lies in the structure of the book itself
My analysis shows that the main mistake of past researchers was seeking the key outside the context. Fable 5.1 made a qualitative leap by focusing on the internal structure of Urquhart's work. The model noticed that the cipher was printed immediately after 32 "procuritations"—numbered wishes by the author, ending with phrases like "this is the wish." Urquhart himself particularly emphasizes the number 32 in the text.
The second clue was an accompanying poem promising the reader that they would find in the cipher their own wishes and the author's thoughts. The decryption logic turned out to be elegant: the i-th number in a line refers to the i-th procuritation, the value indicates the ordinal position of a word within it, and the first letters of those words form the message. The result is a couplet in support of King Charles II: "God support King Charles the Second and make him the supreme ruler of this land." Given that Urquhart was a staunch supporter of the Stuart dynasty, such a finding is entirely consistent with his political stance and is self-verifying: each line contains exactly 32 letters, and the lines rhyme.
A second cryptogram and disputes over authenticity
Using the same method, the model also deciphered the Cyphral Octastich—a more complex eight-line poem of 285 numbers from the treatise The Jewel (1652). Here, the numbers referred to pages of the book. Admittedly, nine letters in the fifth line remained undeciphered due to the absence of the needed leaf in available digitized copies—this would require a physical copy of the edition.
However, not everything is so clear-cut. Specialists in historical ciphers and the Reticuli Labs team have already issued a rebuttal. Their main argument: the algorithm contains at least ten inconsistencies. For example, to obtain the word KING, the letter "K" is required at the 11th position, but in the corresponding procuritation of 80 words, there is not a single one starting with that letter. Exhaustive testing of 65 counting variants yielded only eight matches out of 64 letters—which corresponds to the level of randomness. Furthermore, the original 1653 edition contains no numeric lines at all—they are known only from Urquhart's 1899 biography.
My verdict as an analyst: It is too early to speak of a complete victory of AI over historical ciphers. Although Fable 5.1 demonstrated an impressive capacity for unconventional thinking (it took only 44 minutes and 176,000 tokens), the lack of physical confirmation and logical inconsistencies cast doubt on the purity of the experiment. This is more a demonstration of potential than a definitive solution. Time will tell who is right in this dispute, but the very fact of such a discussion underscores the growing role of AI in solving problems that were once considered unsolvable.