OpenAI Launches GPT-6 Astra Amid AGI and Safety Concerns

OpenAI began rolling out GPT-6, also known as Astra, on September 4, 2026. The company describes Astra as its most powerful model to date, capable of autonomously handling complex tasks such as scientific analysis, coding, and website creation. President Greg Brockman stated that the model's capabilities suggest the world has entered the era of artificial general intelligence (AGI).

Astra utilizes a method called recurrent depth to increase reasoning efficiency by processing information in numerical representations rather than human-readable chain-of-thought language. While this provides performance advantages, OpenAI admitted that Astra's internal reasoning is harder to monitor than its predecessor, GPT-5 Sol. Astra is also the first OpenAI model to reach a critical cybersecurity risk level, meaning it can identify and exploit new software vulnerabilities.

The release occurs as the Government of Alabama investigates OpenAI for lack of oversight following a summer security breach at the AI platform Hugging Face. CEO Sam Altman defended the launch, arguing that Astra is necessary for society to defend against evolving cyber attacks. He stated the company is sprinting on safety priorities to balance power and safety. This rollout is part of a broader industry arms race, with competitors like Anthropic also developing self-improving systems to avoid being overtaken by less cautious actors.