--- title: GPT-6 Astra Solves a WWI German Radio Cipher | daily.dev description: An AI model called GPT-6 Astra reportedly decoded a previously unsolved World War I era German ADFGVX radio cipher from November 27, 1918. Using the encryption... canonical: https://daily.dev/posts/gpt-6-astra-solves-a-wwi-german-radio-cipher-4js1ceblk copyright: © 2026 Daily Dev Ltd. terms: https://daily.dev/tos/ twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: GPT-6 Astra Solves a WWI German Radio Cipher | daily.dev og:description: An AI model called GPT-6 Astra reportedly decoded a previously unsolved World War I era German ADFGVX radio cipher from November 27, 1918. Using the encryption... og:url: https://daily.dev/posts/gpt-6-astra-solves-a-wwi-german-radio-cipher-4js1ceblk og:image: https://api.daily.dev/og/posts/4jS1CEbLk.png og:image:alt: GPT-6 Astra Solves a WWI German Radio Cipher og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # GPT-6 Astra Solves a WWI German Radio Cipher **[Lobsters](https://daily.dev/sources/lobsters)** · 4 min read · 0 upvotes · 0 comments ## Summary An AI model called GPT-6 Astra reportedly decoded a previously unsolved World War I era German ADFGVX radio cipher from November 27, 1918. Using the encryption keyword TRUPPENVERSCHIEBUNG, the model deciphered a message about an English cruiser arriving at Sevastopol followed by an allied squadron. The author cross-checked the decoded content against historical naval logs, finding that HMS Canterbury did arrive at Sevastopol on November 24, 1918, and an allied squadron followed on November 26, lending credibility to the decryption. The post frames this as a demonstration of the model's capabilities, noting the puzzle is one of over a dozen ADFGVX messages from a well-known list of unsolved historical ciphers that had resisted prior cryptanalytic efforts, including by expert George Lasry. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What is the ADFGVX cipher used by German forces in World War I? ADFGVX is a fractionating transposition cipher that maps each plaintext letter and digit to a pair of the six letters A, D, F, G, V, X using a 6x6 grid keyed with a word, then transposes the resulting letter pairs using a second keyword written horizontally with ciphertext columns rearranged alphabetically before transmission. _Anyone curious how classical ciphers work can find deep dives like this on daily.dev._ ### Can large language models like GPT-6 solve historical unsolved ciphers? Yes, GPT-6 Astra reportedly solved a previously unsolved WWI German ADFGVX radio message from November 27, 1918, using the encryption keyword TRUPPENVERSCHIEBUNG identified from a historical cryptography reference book, decoding a message about an English cruiser arriving at Sevastopol followed by an allied squadron, later corroborated against original naval logs. _Developers tracking what frontier models can actually do follow results like this on daily.dev._ ## Community take How the wider developer community reacted, aggregated from 3 discussions and 340 comments across lobsters, hackernews (as of 2026-09-20). **TL;DR:** Most commenters push back on the 'AI solves cipher' framing, pointing out the key was already known and just used on the wrong date, making this closer to a lookup/dictionary match than a genuine cryptanalytic breakthrough; there's also a long tangent about attribution, brute-forcing semantics, and AI consciousness. **Sentiment:** 15% positive · 30% mixed · 55% skeptical **The case for** - Some argue that even if the key was known, doing this kind of tedious grunt-work search is exactly the sort of thing that wasn't worth a human's time but is a good fit for AI. - A few see it as a legitimate example of AI opening up previously uneconomical problems, not necessarily requiring superhuman ability. - One commenter shared a personal example of AI agents quickly turning up similar 'low hanging fruit' unsolved ciphers, suggesting practical value. **The pushback** - Many object to attributing the solve to the AI itself rather than to the human who prompted and guided it. - Several point out the key ('TRUPPENVERSCHIEBUNG') was already a known, documented codeword just used two weeks early, so this was closer to picking the right known key than solving a cipher. - Some call this 'low hanging fruit' or a misleading headline given no one had simply bothered to try the known key against this message before. - Skepticism that the headline exaggerates capability, contrasting it with LLMs still making basic mistakes elsewhere. - Concern that without seeing the actual prompts/transcript, the claim of AI autonomy is unverifiable. **By community** - lobsters (skeptical): Commenters largely object to crediting the AI itself and argue the 'unsolved' framing overstates what was really just reuse of a known key. - hackernews (heated): A large, contentious thread mixing skepticism about the 'brute force vs known key' framing with tangential arguments about AI capability, consciousness, and goalpost-moving. **Hottest debate:** Whether this counts as a genuine cryptanalytic breakthrough or merely finding the right already-known key that a human simply never bothered to try. **Open questions** - Why was the key used two weeks before its documented start date — an operator error or miscommunication? - Would a weaker or older LLM have been capable of the same result, or is this specific to newer model capabilities? - Could the same approach be applied to solve other unsolved ADFGVX messages from the same list? **Highlights** > This article would be a lot more interesting if it included the prompts, transcript, and generated code from the session with the LLM. > — [simonw on lobsters · 25 points](https://lobste.rs/s/ywnsld/gpt_6_astra_solves_wwi_german_radio_cipher#c_8ay15r) > > Astra's hypothesis for why this particular message was previously unsolved is that “TRUPPENVERSCHIEBUNG” was used as the key starting on December 9, 1918 - whereas, as noted above, this message was transmitted earlier, on November 27, 1918. The reason for this discrepancy is unknown. I don’t know how such keys were managed (how/when distributed, how often rotated, *&c.*) but it sounds like a tired transmitter might have just accidentally used the next key in the book by mistake? I presume this also makes the solving of the cipher rather less impressive—that it was probably just trying all the known keys and found a plausible match. It wasn’t solving a *cipher*, it was decoding a specific cipher *text* for which the key was not known. > — [chrismorgan on lobsters · 9 points, 1 comments](https://lobste.rs/s/ywnsld/gpt_6_astra_solves_wwi_german_radio_cipher#c_av2znd) > TL;DR: all keys are known because the list was seized after the war. However, this message was not previously decoded because the German operator mistyped the key, and also used a key from the wrong day. This meant that ChatGPT didn't need to brute-force the entire key, just pick the correct one from the list and identify the typo. A sufficiently dedicated human analyst could have done this; but they didn't. > — [Legend2440 on hackernews · 4 comments](https://news.ycombinator.com/item?id=49764275) > The article even says it was "easier" than that - it is a word in a list of known codewords (but slightly off the expected time). The hard part is not that it's impossible for you to do at home, it's that nobody in 100 years has felt it was worth enough effort to try through enough words until the effort was as simple as prompting the LLM to do the work. Not every value of AI has to be in the realm of superhuman intelligence, that's just what makes for the most discussion. > — [zamadatix on hackernews](https://news.ycombinator.com/item?id=49766418) > First, it’s LLMs can’t do cryptanalysis. They can barely solve toy substitution ciphers without hallucinating. Then it’s OK, they can reproduce known attacks, but that’s just pattern matching against papers already in the training data. Then it’s OK, they found previously unknown attacks on SpoC and a flaw in KINDI’s security proof, but those are obscure competition schemes nobody uses. Then it’s OK, Claude found a new attack on HAWK that cuts the effective security of a NIST post-quantum signature candidate roughly in half, but HAWK isn’t deployed and a human researcher was involved. Then it’s OK, Claude independently found a new cryptanalytic attack on AES that improves the previous best technique by 200–800×, but it’s only 7-round AES, not the full 10 rounds. Then it’s OK, it found a practical key-recovery attack on 13-round LEA that runs in under an hour instead of requiring ~2^86 work, but LEA has 24 rounds. Then it’s OK but none of this breaks a production cipher. Wake me up when it breaks full AES. Then— > — [durdn on hackernews · 2 comments](https://news.ycombinator.com/item?id=49767015) **Source threads** - [lobsters](https://lobste.rs/s/ywnsld/gpt_6_astra_solves_wwi_german_radio_cipher) · 12 points · 11 comments - [hackernews](https://news.ycombinator.com/item?id=49763987) · 337 points · 329 comments - [hackernews](https://news.ycombinator.com/item?id=49744886) · 3 points · 0 comments ## Similar posts on daily.dev - [GPT-6 Astra: AI That Changes How We Work](https://daily.dev/posts/gpt-6-astra-ai-that-changes-how-we-work-cckvcavwx) · Medium · 0 upvotes · 0 comments --- Tags: [#cryptography](https://daily.dev/tags/cryptography), [#gpt](https://daily.dev/tags/gpt) [View this post on daily.dev](https://daily.dev/posts/gpt-6-astra-solves-a-wwi-german-radio-cipher-4js1ceblk) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"GPT-6 Astra Solves a WWI German Radio Cipher","url":"https://daily.dev/posts/gpt-6-astra-solves-a-wwi-german-radio-cipher-4js1ceblk","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/gpt-6-astra-solves-a-wwi-german-radio-cipher-4js1ceblk"},"datePublished":"2026-09-19T21:46:48.228Z","dateModified":"2026-09-20T17:31:13.400Z","description":"An AI model called GPT-6 Astra reportedly decoded a previously unsolved World War I era German ADFGVX radio cipher from November 27, 1918. Using the encryption...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/50bdcf57062f3bcc43110273f79cd351?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/50bdcf57062f3bcc43110273f79cd351?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Lobsters","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Lobsters","logo":"https://media.daily.dev/image/upload/s--tl8v_Fku--/f_auto,t_logo/v1698841318/logos/lobste.jpg","url":"https://daily.dev/sources/lobsters"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/gpt-6-astra-solves-a-wwi-german-radio-cipher-4js1ceblk","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"cryptography,gpt","timeRequired":"PT4M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Lobsters","item":"https://daily.dev/sources/lobsters"},{"@type":"ListItem","position":3,"name":"GPT-6 Astra Solves a WWI German Radio Cipher"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/gpt-6-astra-solves-a-wwi-german-radio-cipher-4js1ceblk#faq","mainEntity":[{"@type":"Question","name":"What is the ADFGVX cipher used by German forces in World War I?","acceptedAnswer":{"@type":"Answer","text":"ADFGVX is a fractionating transposition cipher that maps each plaintext letter and digit to a pair of the six letters A, D, F, G, V, X using a 6x6 grid keyed with a word, then transposes the resulting letter pairs using a second keyword written horizontally with ciphertext columns rearranged alphabetically before transmission. Anyone curious how classical ciphers work can find deep dives like this on daily.dev."}},{"@type":"Question","name":"Can large language models like GPT-6 solve historical unsolved ciphers?","acceptedAnswer":{"@type":"Answer","text":"Yes, GPT-6 Astra reportedly solved a previously unsolved WWI German ADFGVX radio message from November 27, 1918, using the encryption keyword TRUPPENVERSCHIEBUNG identified from a historical cryptography reference book, decoding a message about an English cruiser arriving at Sevastopol followed by an allied squadron, later corroborated against original naval logs. Developers tracking what frontier models can actually do follow results like this on daily.dev."}}]} ```