OpenAI Calls Off GPT-6.1 Astra Launch, Details Safety Cases for Frontier Training

OpenAI has canceled the launch of its GPT-6.1 Astra model, which was initially planned for release in October. The model was intended to be integrated into ChatGPT and Codex, but it did not meet expectations. OpenAI is focusing on safety cases for frontier training, indicating a prioritization of safety in its development process. The cancellation suggests that the company is taking a cautious approach to ensure the reliability and security of its models. The decision to postpone the launch may allow OpenAI to further refine and test the GPT-6.1 Astra model before its release.