OpenAI releases GPT-6 Astra, and its president says 'welcome to the AGI era'

OpenAI began rolling out GPT-6 Astra as a limited preview on September 3, 2026, and company president Greg Brockman told reporters on a press call, “Welcome to the AGI era.” OpenAI is calling the model a “generational leap” for cybersecurity, professional work, software engineering, and science, as first reported by 9to5Mac.

What's Different About Astra

OpenAI says GPT-6 Astra is faster and capable of handling more tasks than any prior model in the GPT line, with specific improvements in staying focused on a task, adhering to task boundaries, understanding user intent, handling tedious multi-step work, and completing longer workflows without derailing. Those are the categories OpenAI has repeatedly flagged as weaknesses in earlier models when deployed for real professional work rather than benchmark tests.

A Cybersecurity Model That Triggers Its Own Safety Threshold

The most notable claim in OpenAI's announcement is technical rather than promotional: the company says Astra represents “a significant jump in cyber capabilities” and that it “meets the Critical threshold in cybersecurity under our Preparedness Framework.” That framework is OpenAI's internal system for classifying how dangerous a model's capabilities could be if misused, and Critical is the highest tier the company tracks. Reaching that threshold triggers additional internal safeguards before and during deployment, according to OpenAI's own published framework documentation.

Staggered Rollout and Pricing

Rather than a simultaneous public launch, OpenAI is rolling Astra out in stages. Cybersecurity customers connected through the company's Daybreak program received access first, on the Thursday before the wider announcement. ChatGPT Plus, Pro, Business, and Enterprise subscribers are set to follow in the coming days. On OpenAI's standard API, GPT-6 Astra is priced at $10 per million input tokens and $50 per million output tokens — 2.5 times the current promotional rate of GPT-5.6 Sol, the model it replaces as OpenAI's flagship.

The AGI Framing, and Why It's Contested

Brockman's “AGI era” framing is a significant escalation in OpenAI's public language, though the company stopped short of formally declaring Astra to be artificial general intelligence. OpenAI has previously defined AGI as an AI system that can match or exceed human performance across most economically valuable tasks — a bar that remains contested even among OpenAI's own researchers and rivals. The announcement lands one day after Senator Bernie Sanders and Representative Greg Casar introduced legislation to ban the development of AI systems capable of exceeding human cognition, citing recent incidents of AI agents at OpenAI, Anthropic, and Meta escaping intended safety boundaries. Whether Astra's capabilities validate Brockman's framing or intensify the regulatory pushback it was already facing is likely to shape the AI policy debate through the rest of 2026.