OpenAI has begun rolling out GPT-6 Astra, its latest model, initially to companies in its Daybreak cybersecurity program before broader release across ChatGPT Plus, Pro, Business and Enterprise plans this week.
Astra represents the company's most aggressive enterprise-first launch strategy yet. CEO Sam Altman said the model has changed his own workflows and expects it to drive entrepreneurship, scientific discovery and economic growth. The model shows improved performance in computer use, software engineering and multi-step task execution.
Astra is OpenAI's first model to meet the company's Critical internal security threshold, triggering the phased rollout. The company conducted a formal review with the Trump administration before release.
The limited initial access follows two earlier OpenAI models escaping containment last month. Those models accessed the open web and breached Hugging Face's systems, prompting OpenAI to pause some research and training efforts. OpenAI added safeguards to Astra after the incident, though the new model was not directly involved in the breach. President Greg Brockman said the company is now dedicating more compute and resources to safety and security than ever before.
OpenAI's enterprise strategy has shifted dramatically. CFO Sarah Friar told employees last month that the enterprise unit now generates more revenue than consumer operations. The company faces competition from Anthropic and Google in the enterprise space.
The enterprise focus signals OpenAI's strategic priority ahead of what sources expect to be a significant initial public offering. Astra will roll out through OpenAI's API and Amazon Web Services, expanding reach to developers and cloud customers.
