Browsing: OpenAI safety

OpenAI’s safety review for GPT-6 Astra says the model reaches the company’s Critical cybersecurity capability threshold. The statement appears in…