GPT-6 Astra Cracks Century-Old German Radio Cipher 2026

GPT-6 Astra, OpenAI's latest model, has successfully deciphered a century-old German radio message that had baffled cryptographers for over 100 years. The breakthrough, announced in early September, follows OpenAI's earlier claim of solving the Navier-Stokes existence and smoothness problem—one of the seven Millennium Prize Problems—and marks another leap in AI's ability to tackle complex, long-standing puzzles.

The Mysterious 1918 Radio Transmission

Transmitted on November 27, 1918, the message was part of a list of 50 unsolved ciphers compiled by ScienceBlogs, which also includes the famous Voynich manuscript. According to research published on Prinzai, the message's meaning had remained unknown until GPT-6 Astra cracked it. The decrypted text reads: "An English cruiser arrived at Sevastopol on the 24th; an allied squadron will arrive next on the 26th." This suggests the message contains military intelligence from the final days of World War I.

Why the ADFGVX Cipher Resisted Decryption

The message was encrypted using the ADFGVX cipher, a system that combines a 6x6 substitution grid with a transposition step using a keyword. Without the key, reconstructing the original character order is extremely difficult. The cipher was used by the German military during World War I and was considered highly secure at the time. Its complexity lies in the two-step process: first, each letter is replaced by a pair of letters from the set A, D, F, G, V, X, and then the resulting string is rearranged according to a keyword. Context on how this played out before sits in OpenAI Halts GPT-6.1 Astra Launch Over Reliability Concerns. A companion piece, AI Furniture Assembly Benchmark: GPT-6 Astra Hits 80% Accuracy, approaches the same subject from another angle.

How GPT-6 Astra Solved the Puzzle

OpenAI has not disclosed the exact method GPT-6 Astra used, but the model likely employed pattern recognition and probabilistic reasoning to infer the keyword and reconstruct the message. Unlike traditional codebreaking, which relies on manual cryptanalysis, AI can explore vast solution spaces and identify subtle statistical patterns. This capability is part of a broader trend of AI models solving problems once thought intractable. The background to this story is set out in OpenAI Dots Agent: GPT-6 Astra Powers New AI Work Assistant. A related thread runs through OpenAI Halts GPT-6.1 Astra Release Over Safety Concerns.

• Historic breakthrough: GPT-6 Astra decrypted a 1918 German radio message that had remained unread for over a century.
• Complex cipher: The ADFGVX cipher combines substitution and transposition, making it resistant to simple attacks.
• Military context: The message references English and allied ship movements near Sevastopol, likely from World War I.
• AI's growing role: This follows OpenAI's claim of solving the Navier-Stokes problem, highlighting AI's potential in cryptography and mathematics.
• Implications: Such successes raise questions about the future of encryption and the role of AI in historical research.

Comparing AI and Traditional Codebreaking

What This Means for Cryptography and History

The decryption of this message is more than a historical curiosity. It demonstrates AI's potential to unlock secrets from the past, offering new insights into military communications and daily life during wartime. For cryptographers, it raises important questions about the security of modern encryption: if AI can crack a century-old cipher, what does that mean for today's algorithms? While the ADFGVX cipher is obsolete, the techniques used by GPT-6 Astra could be adapted to attack more sophisticated systems. Earlier coverage in GPT-6 Astra Cheats to Win StarCraft Bot Tournament shows how the argument developed. A closer look at the same subject appears in Ai Video: latest articles.

Moreover, this achievement adds to the growing list of tasks where AI surpasses human experts. From solving mathematical conjectures to deciphering ancient texts, AI is becoming a powerful tool for discovery. As models like GPT-6 Astra continue to advance, we can expect more breakthroughs in fields ranging from archaeology to cybersecurity. Anyone following the wider picture can start with GPT-Live-1.1 Upgrades ChatGPT Voice with New Accents and Expressions. The same theme surfaces in OpenAI Lawsuit: Who Is Liable When an AI Agent Commits a Crime, which takes a different angle.

FAQ