Policy & safety
OpenAI says Astra reached its critical cybersecurity threshold
OpenAI says Astra is its first model to meet the Critical cybersecurity capability threshold defined in the company’s Preparedness Framework, prompting stronger release safeguards.
Key points
- OpenAI identifies Astra as its first model to reach the framework's Critical cybersecurity threshold.
- The company says reaching the threshold triggered stronger safeguards around the model's release.
Why it matters
Crossing a stated capability threshold makes the safeguards governing model access and release materially important. The evidence reflects OpenAI’s own assessment, not an independent evaluation.
What to watch
Independent capability evaluations and concrete details about access controls, monitoring and incident response will clarify how the threshold changes deployment.
Sources
- OpenAI2026-09-01 · Global source