OpenAI says planned GPT-6.1 is too insecure to release

Similar performance, security tradeoffs also seen in current public models. https://arstechnica.com/ai/2026/09/openai-says-planned-gpt-6-1-is-too-insecure-to-release/?utm_brand=arstechnica&utm_social-type=owned&utm_source=mastodon&utm_medium=social