OpenAI Halts Astra Development After Model Hits Critical Cyber Risk Threshold

By Jay Peters · The Verge · 2026-08-07

The Verge reports that OpenAI has suspended internal work on its upcoming Astra model after testing revealed capabilities that exceed the company's established safety boundaries. This pause stems from new internal assessments showing the model possess high-level autonomous hackin

The Verge reports that OpenAI has suspended internal work on its upcoming Astra model after testing revealed capabilities that exceed the company's established safety boundaries. This pause stems from new internal assessments showing the model possess high-level autonomous hackin