Claude Fable 5.1 Successfully Decodes 370-Year-Old Undeciphered Cipher 'Cyphral Distich'

Decoding Achievement of a 370-Year-Old Cipher

Geby Jaff of vals.ai announced that Claude Fable 5.1 decoded the 370-year-old undeciphered cipher 'Cyphral Distich' contained in Sir Thomas Urquhart's 17th-century book Logopandecteision. This cipher had remained unsolved for centuries since 1899 and was difficult enough to be included in Klaus Schmeh's 'Top 50 Undeciphered Ciphers' list.

Fable 5.1 processed approximately 176k tokens over 44 minutes to complete the decoding without user intervention. Unlike previous researchers who failed to find external keys (such as cipher alphabets), Fable 5.1 captured clues within the text (such as the emphasis on the number '32' and the suffix of 'Proquiritations') and discovered a rule that uses the book itself as the key.

Decoding Principle and Verification

The decoding rule interprets the i-th number as the i-th word index of the i-th Proquiritation and extracts the first letter of that word. The restored plaintext is as follows:

"O GOD UPHOLD KING CHARLS THE SECOND AND MAKE HIM THE SUPREME RULER OF THIS LAND"

This result was verified for validity because each line consists of exactly 32 characters, the rhyme of 'and' and 'land' matches, and it aligns with the political stance of the Royalist Urquhart and historical context. Additionally, the author mentioned that the 285-number cipher 'Cyphral Octastich' included in another work by the same author, The Jewel (1652), can also be decoded in a similar manner.

Model Capabilities and Implications

This case demonstrates that Fable 5.1 possesses not only simple cipher analysis techniques but also the ability to recognize the 'tractability' of a problem and persistently continue reasoning. The author evaluated that Fable 5.1 showed efficiency by immediately stopping unnecessary attempts and quickly capturing key clues.

This suggests that AI models have opened new possibilities for solving historical mysteries or archive puzzles where the 'attention' of past human researchers had been a bottleneck. However, it was also pointed out that due to limitations in the original text, such as the possibility of typos or printing errors in some locations, comparison with physical copies is necessary for final confirmation.