The Race — Building Democracy in the Age of AI. People race toward a shared finish line.
The Race
← Specific Events

Global

Scorecard

OpenAI designates Astra Critical in cybersecurity and activates differentiated frontier access

What happened

OpenAI formally determined that Astra meets the Critical cybersecurity threshold under its Preparedness Framework, the first OpenAI model to receive that designation. Astra demonstrated autonomous vulnerability discovery and exploit development, including a two-zero-day exploit chain. OpenAI plans broad release with more restrictive default safeguards while reserving advanced cyber capability initially for a small tester group and later expanding access through Daybreak Blue.

Potential direction of power

Concentration · Concentration +1

Astra converts differentiated frontier access from a prospective safety mechanism into an operating allocation regime. Access to the model and access to its strongest capabilities are now materially different things, and OpenAI controls qualification, monitoring, expansion, and intervention. The score stops at +1 because the trusted-user structure also distributes defensive capability across a broader cyber ecosystem rather than reserving it solely to OpenAI or government.

Window of contestability

Narrows

Provider-controlled qualification and safeguards narrow access to the strongest cyber capabilities, with no independent adjudication mechanism for access decisions.

Supporting sources

  1. openai.com · Supporting source 1 ↗