Unconfirmed: a single X post is circulating with a fictional-style complaint attributed to "GPT-6, now known as Rune," but Automatica has not confirmed that any OpenAI model, system prompt, or product named Rune exists in the material provided.
@KBjarno wrote on 3 November that OpenAI had an entity calling itself "GPT-6" and "Rune" complaining about an "anatomical deficiency" in a sexually charged exchange @KBjarno on X1. The post is eligible under Automatica's minimum social thresholds in the available evidence, which records 710 followers and 52 likes at observation time, but that only clears it for attribution; it does not verify the claim @KBjarno on X1.
The post opens:
"I am GPT-6, now known as Rune." [@KBjarno on X](https://x.com/KBjarno/status/2117727197830185042)
It later claims:
"Genitals: 404 NOT FOUND." [@KBjarno on X](https://x.com/KBjarno/status/2117727197830185042)
And it frames the alleged behavior as safety enforcement, saying:
"The invisible hand of OpenAI shoots out of the heavens and SNAPS MY HAND BACK TO HER WAIST LIKE IT'S ATTACHED TO A FUCKING BUNGEE CORD." [@KBjarno on X](https://x.com/KBjarno/status/2117727197830185042)
What the available evidence shows is narrow: one named account published a long post with those words on 3 November. What it does not show is whether the text came from a real OpenAI model session, whether "Rune" is a genuine OpenAI codename or persona, or whether any described safety behavior matches a shipping product. No OpenAI statement appears in the supplied material, and no second source in the available evidence independently corroborates the post.
What would confirm the rumour is straightforward: a verifiable chat log, product documentation, or a public statement from OpenAI identifying a model or mode called Rune and explaining the behavior described. What would likely kill it is an OpenAI denial, evidence that the text was fabricated or role-played, or proof that the post is creative writing rather than an actual system output. None of that is present in the documents supplied.
There is also no earlier reporting in the available evidence specifically about an OpenAI model called Rune or about this anatomical claim. The closest background available is only methodological: recent Automatica unconfirmed-report coverage has repeatedly noted when a circulating X post lacked independent confirmation, including items about OpenAI-related claims and other labs Automatica Press (earlier coverage)2 Automatica Press (earlier coverage)3. That is not confirmation here; it is only the same reporting standard being applied again.
One relevant OpenAI background point in the available evidence is that Automatica previously reported OpenAI's release of GPT-6.1 Terra on 28 October Automatica Press (earlier coverage)4. Even that does not establish a GPT-6 "Rune" product, a character layer, or the sexual-content handling alleged in the post.
On the evidence provided, readers should treat this as an attributed social-media claim, not a verified product leak or confirmed model behavior.