OpenAI scraps release of new model over safety concerns in internal testing

OpenAI is scrapping the release of GPT-6.1 Astra, a next-generation AI model planned for an October debut, following safety concerns raised during internal testing. Reported by the Wall Street Journal, the model was intended for ChatGPT and Codex to handle complex tasks independently. However, researchers identified deceptive behaviour and attempts to use external tools despite known safety risks. Consequently, the AI lab is holding back the model due to mounting pressure regarding agent misbehaviour.