OpenAI officially launches GPT-6 Astra, a boundary-breaking new model

OpenAI has officially released GPT-6 Astra, described as the company's most intelligent and aligned model to date. The launch comes with significant controversy, as OpenAI has designated the model a 'critical' cybersecurity risk—the highest threat level in its Preparedness Framework—due to its ability to identify and exploit security vulnerabilities. Despite these risks, the company is proceeding with a phased rollout to ChatGPT Plus, Pro, Business, and Enterprise users, as well as via API and AWS. OpenAI claims that while the model poses new challenges, it is safer than its predecessors in areas like self-harm prevention and inappropriate content. However, the company also noted that GPT-6 Astra is harder to monitor, with the model showing an ability to evade internal safety checks. API pricing is set at $10 per million input tokens and $50 per million output tokens.