← Arethusa

KRYPTOS

Sanborn sculpture · K1–K4

Between subtle shading and the absence of light lies the nuance of iqlusion. A working bench for the four passages — decode the solved three live, and test your hand against K4.

SECTIONS K1 K2 K3 SOLVED · K4 OPEN · 97 CHARACTERS UNBROKEN

K1

Solved · keyed Vigenère

A polyalphabetic Quagmire 3 cipher built on a keyed alphabet KRYPTOSABCDEFGHIJLMNQUVWXZ, driven by the keyword PALIMPSEST. Change either field and watch the plaintext re-derive in real time.

Ciphertext62 chars
Live plaintexthover a column below to trace it in the tableau
Keyed Vigenère tableau ▾

K2

Solved · keyed Vigenère

Same machinery and same keyed alphabet as K1 — only the keyword changes to ABSCISSA. The decoded message hides geographic coordinates and a pointer to "layer two."

Ciphertext372 chars
Live plaintext
Keyed Vigenère tableau ▾

K3

Solved · double transposition

No substitution — every letter is present, only reordered. The plaintext is regridded and rotated twice: into a 42 × 8 grid read by columns (a 90° turn), then into a 14 × 24 grid and turned again. Press decode to run the verified inverse.

Ciphertext336 chars
Recovered plaintext

K4

Open · unbroken since 1990

97 characters, an unknown method. Type your candidate plaintext into the cells beneath each ciphertext letter. The four spans Sanborn has publicly confirmed are pre-locked in gold — any real method must reproduce them exactly, so the bench checks your guesses against them as you type.

confirmed clue (locked) matches the clue conflicts with a clue
PositionsCiphertextPlaintextReleased
22–25FLRVEAST2020
26–34QQPRNGKSSNORTHEAST2020
64–69NYPVTTBERLIN2010
70–74MZFPKCLOCK2014
Frequency analysis · 97 characters
Index of coincidence
Distinct letters used
Most frequent
Repeated bigrams

Letter distribution
Repeated sequences & gaps — even gaps can betray a key period (Kasiski)

Where the wall is

K4 is not the simple Quagmire of K1/K2 — a plain keyed Vigenère will not reproduce the locked spans, which is your fastest filter for a dead theory. Sanborn says the solutions to K1–K3 (the keyed alphabet, the planted misspellings, the back-panel tableau) all feed K4. "BERLIN CLOCK" likely points beyond plaintext toward a decoding step or an external referent — the lamp-based Berlin clock fits the work's running theme of light.

On "solutions" you'll find online

Treat every claimed full K4 break with suspicion. In 2025 two researchers found the genuine plaintext inside Sanborn's own papers at the Smithsonian — not by codebreaking. Those archives are now sealed to ~2075, the finders won't publish, and the complete archive (with a further passage Sanborn calls K5) sold at auction for $962,500. No public, reproducible method exists, and the real text is withheld — so no third party can validate a claimed solution against it. Build from the confirmed clues above, not from someone's reveal.