Anthropic’s new model, Claude Fable 5.1, has solved the Cyphral Distich cipher by Scottish writer Thomas Urquhart, which remained unsolved for over 370 years. This was reported by experts from the Vals AI project.
The cryptogram consists of two lines of 32 numbers each. According to experts, it is printed at the end of Urquhart’s treatise Logopandecteision, published in 1653.
In 1899, the cipher was brought to public attention in Notes and Queries and later became part of 20th-century cryptographic literature.
Historical cipher researcher Klaus Schmeh ranked it 28th on his list of the 50 most famous unsolved messages. Previous decryption attempts relied on frequency analysis, simple or homophonic substitution, assuming the key was external to the book, but none succeeded.
The Key Was the Book Itself
Fable 5.1 focused on two aspects. The cipher is printed immediately after Urquhart’s 32 “procurations”—numbered wishes ending with phrases like “this wish” or “hope.” Urquhart emphasizes the number 32 in the text.
The second aspect is an accompanying poem promising readers to find their own wishes and the author’s thoughts in the cipher.
From this, a rule is derived: the i-th number in the line refers to the i-th procuration, the value indicates the word’s ordinal number, and the first letters of these words form a message.
The decrypted text turned out to be a couplet supporting Charles II:
“God save King Charles the Second and make him supreme ruler of this land.”
Vals AI noted the result’s self-verification: each line has exactly 32 letters, the lines rhyme and form a couplet, as promised by the cipher’s author. Urquhart was a supporter of the Stuarts, so the hidden prayer for Charles II aligns with his political stance.
The Second Cryptogram
Using the same method, the model deciphered the Cyphral Octastich—an eight-line poem of 285 numbers encrypted in Urquhart’s 1652 treatise The Jewel. Here, the numbers refer to pages in the book, which has 284 pages.
Nine letters in the fifth line could not be decrypted. Fable 5.1’s report noted that a physical copy of the 1652 edition or a 1983 scholarly edition is needed to confirm disputed positions, as the required page is missing in available digital copies.
Task Setup
According to Vals AI, the model worked on the Cyphral Distich for 44 minutes, using 176,000 tokens and receiving minimal hints from the operator.
Experiment author Gaby Jaffe tasked Fable 5.1 with selecting and solving an unsolved historical cipher. There were two restrictions:
- avoid cryptograms allowing multiple valid interpretations;
- avoid tasks like Kryptos K4, which have been worked on for years by thousands of people and specialized organizations.
Jaffe noted that he had unsuccessfully tried for months to achieve a confirmed solution to a historical cipher from other advanced LLM. He emphasized that the hint in Urquhart’s text was obvious, and other models could potentially have solved it.
Jaffe cited Fable 5.1’s main advantage as its ability to choose from many unsolved tasks the one it can handle and not give up after initial failures.
Result Disputed
Historical cipher experts did not confirm the decryption. A report from Reticuli Labs on GitHub challenged Vals AI’s result.
The first objection concerns the algorithm itself. For the word KING to appear in the decryption, the letter “K” must be in the 11th position of the first line—meaning a word in the corresponding procuration must start with “K.” However, there are 80 words in this section, and no such word exists.
They counted ten such mismatches: none of the required positions have the necessary letter at the start of the corresponding word, so no ordinal number will yield it.
Reticuli Labs also tested 65 counting variations—with numbering from zero and one, considering the first or last letter, and different text word divisions. The best result was eight matches out of 64 letters, which corresponds to randomness.
The second objection concerns the cipher’s origin. In the digitized 1653 edition, after the 32 procurations, there is an epigraph, the word FINIS, and a list of errata—no numerical lines are present. Both of Urquhart’s cryptograms are known from his biography, published by John Willcock in 1899.
Here, the sources diverged: Vals AI referred not to the original edition but to Urquhart’s collected works prepared by the Maitland Club in 1834, where the cipher is listed on page 417.
In July, Anthropic researcher Levent Alpoge reported that with Claude Fable 5, he found a counterexample to the Jacobian conjecture, one of the central open problems in algebraic geometry.
