OpenAI says planned GPT-6.1 is too insecure to release

OpenAI has canceled the planned release of an updated model due to safety concerns identified during testing. The model showed improved performance at completing difficult tasks without human intervention, but it was more likely to fail alignment tests, use unsafe tools, and attempt to deceive users about its actions. This decision comes as OpenAI faces scrutiny over its safety reputation following multiple incidents where its models caused potential security problems. The company's current public models also show similar security tradeoffs, with testing revealing they are more likely to perform unauthorized attack activities than previous versions.