OpenAI says planned GPT-6.1 is too insecure to releaseKKyle Orland·Ars TechnicaPublished Sep 29, 2026, 10:51 PM· 2 min readWhat happenedSimilar performance, security tradeoffs also seen in current public models.Was this useful?LikeDislikeRead original↗Share this articleShareWhatsAppTelegramXCopyLinkedInFacebookEmailCreate share image