OpenAI flags Astra AI model as having critical cyber abilities

OpenAI has a new AI model that can apparently hack things on its own. And for the first time in the company’s history, it’s hitting the brakes before shipping.

The company announced on August 7 that internal evaluations of its upcoming model, called Astra, suggest it has reached “Critical” cybersecurity capabilities under OpenAI’s own Preparedness Framework. That framework, introduced back in December 2023, was designed to categorize exactly this kind of risk. Astra is the first model to trigger its highest alarm level for cyber skills.

What “Critical” actually means

OpenAI’s Preparedness Framework is essentially a traffic light system for dangerous AI capabilities. Models get rated across several risk domains, from “Low” to “Critical.” GPT-5.6, the company’s previous frontier model, earned only a “High” rating in cyber capabilities. Astra apparently blew past that threshold.

Under the framework’s definitions, “Critical” means the model can autonomously identify and develop functional zero-day exploits in hardened systems. The “Critical” designation also means Astra can execute comprehensive cyberattacks starting from nothing more than high-level strategic goals.

OpenAI responded by pausing certain internal development work on Astra. The company implemented heightened security controls, including isolated testing environments and universal monitoring across Astra-related systems. Select partners will reportedly receive early access to the model so they have time to shore up their defenses before any broader release.

A model built differently

Astra is a distinct model from OpenAI’s GPT line, and its capabilities extend well beyond cybersecurity. The model has already demonstrated exceptional mathematical abilities, reportedly solving ten longstanding mathematical problems at a compute cost of roughly $2,000.

OpenAI has been careful to distance Astra from a separate security incident in July 2026 involving Hugging Face, clarifying that Astra played no role in that event.

As of early September 2026, Astra remains unreleased. There’s no public API access, no pricing information, and no confirmed launch date. The company has indicated it eventually aims to deploy Astra’s sophisticated capabilities for defensive purposes.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.