"Key Takeaways"
- Vals AI reported that Claude Fable 5.1 solved Sir Thomas Urquhart's Cyphral Distich in 44 minutes, with no human interjections during the run.
- The reported plaintext is a royalist couplet. The method treats the book itself as a book cipher keyed to Urquhart's 32 Proquiritations.
- Independent challenge work, including Reticuli's write-up and Schneier's note on the dispute, argues the Distich is missing from the examined 1653 printing, and that the method fails against that text.
- The Distich does appear in the 1834 Maitland Club reprint Vals cited. Until the page history is settled, treat the headline as a claim under review.
Summarize this article with
The feed said Claude Fable 5.1 solved a 370-year-old cipher. That sentence is too clean for the evidence.
What we actually have is a strong reported result, a compelling plaintext, a serious replication challenge, and a printing history that is not finished arguing with itself. The useful skill is not picking a team. It is learning how to read an AI discovery claim.
What the Cyphral Distich is
The Cyphral Distich is a short numerical puzzle linked to Sir Thomas Urquhart, the seventeenth-century Scottish writer behind Logopandecteision. In the version discussed online, it is two lines of 32 numbers each. Cryptologist Klaus Schmeh has listed it among the better-known unsolved encrypted messages, which is why the story travelled so far once an AI lab said it had cracked it.
The interesting part of the puzzle, if the reported solve is right, is not exotic mathematics. It is a book cipher idea: each number points into the surrounding text, and the first letters of those words spell a message.
What Vals AI reported Claude Fable 5.1 did
On August 31, 2026, evaluation firm Vals AI published a write-up by Geby Jaff. According to that account, Claude Fable 5.1 was given an open task: solve Urquhart's Cyphral Distich. After 44 minutes, about 176,000 tokens, and what Jaff called zero interjections from him, the model returned a method and a plaintext.
Vals reported the couplet as: "O GOD UPHOLD KING CHARLS THE SECOND AND / MAKE HIM THE SUPREME RULER OF THIS LAND." The spelling CHARLS is period-flavoured, the lines rhyme, and each line is said to resolve to 32 letters. That self-checking structure is why the plaintext looked so persuasive.
Vals also said months of earlier frontier-model attempts had not produced a verified solve, and that Fable 5.1 found the Distich after being steered away from harder targets such as Kryptos K4. Those are Vals' claims about its own testing process, not an independent audit of every lab.
Reported claim, not settled fact
The viral version of this story often compresses three different things into one: that Claude Fable 5.1 produced a coherent decrypt, that the decrypt matches the intended 1653 cipher, and that the original page has been produced and checked. Only the first of those is clearly documented in the Vals write-up.
What the challenge found
Within a day, challenge work pushed back. Reticuli Labs published findings arguing that the Distich does not appear at the end of the examined 1653 printing. In their account, the book closes with Proquiritations 30 to 32, ornaments, a Latin epigraph, FINIS, and errata. No Distich. Schneier on Security linked the dispute and pointed readers at that GitHub write-up.
The challenge also tested the book-cipher method against the 1653 text. According to Reticuli, ten positions are impossible because the required letter never appears as a word-initial in the relevant petition. Across many counting conventions, they reported chance-level recovery. If that holds for the edition being tested, the method does not unlock that printing.
There is a second complication that keeps the story from collapsing into a simple gotcha. Vals cited the Distich inside the 1834 Maitland Club edition of Urquhart's works. Later write-ups, including Wren's Cipher Room, confirmed that page 417 of that reprint does contain THE CYPHRAL DISTICH where the 1653 film instead shows ornaments. So one honest reading is: Fable may have solved a real puzzle in a later printed text, while the 1653 page evidence is still contested.
What can actually be established right now
- Vals AI published a detailed solve claim for Claude Fable 5.1, including runtime, token count, method, and plaintext.
- The reported plaintext is internally tidy: length, rhyme, and politics all line up with a royalist couplet.
- Independent challenge work says the Distich is absent from the examined 1653 printing, and that the method fails there.
- The Distich does appear in the 1834 Maitland Club reprint that Vals cited.
- Neither side has closed the case with a universally accepted original-page artifact that ends the printing debate.
What this teaches about impressive AI discoveries
A fluent model can look like a historian, a cryptanalyst, and a librarian at once. That is useful. It is also why impressive AI discoveries need the same boring checks as any other discovery: what source was used, what page was shown, who replicated the method, and what would falsify the claim.
Claude Fable 5.1 may still have done something genuinely interesting here. Persistence across a long unsolved puzzle is not nothing. But "solved a 370-year-old cipher" is a magazine headline. The careful version is: an evaluation firm reported an autonomous solve against a reprint that contains the Distich, while challenge researchers say the 1653 printing they inspected does not support the same story.
If you only remember one thing from the Cyphral Distich week, make it this. A striking AI result is not automatically a verified result.
Frequently Asked Questions
Did Claude Fable 5.1 solve the Cyphral Distich?
Vals AI reported that it did, in 44 minutes, with a coherent royalist plaintext. Independent challenge work disputes whether that Distich exists in the examined 1653 printing and whether the method works there. The claim remains under review.
What is an AI solved 370 year old cipher claim really saying?
Usually three claims at once: the model produced a decrypt, the decrypt matches the historical puzzle, and the historical page has been checked. Ask for each piece separately.
Why does the Claude Fable cipher story still matter?
Because long autonomous reasoning is getting better, and because public AI claims now travel faster than document checks. The Distich fight is a useful rehearsal for reading the next big discovery headline.
Should I trust viral AI history breakthroughs?
Treat them as leads. Look for the primary source page, independent replication, and clear labels for what is confirmed versus what is still contested.
The Chai Takeaway
Keep the wonder. Keep the footnotes. When an AI model looks like it just closed a centuries-old case file, ask which page it read, who checked that page, and what would prove the story wrong. That habit will age better than any one cipher headline.
Explore More AI Updates