Quotes on mitigating biological and chemical AI risks

We have a proactive multi-layered defense stack which includes model safety training.
Unknown[1]
These safeguards sufficiently minimize the associated risks under our Preparedness Framework.
Unknown[1]
We believe this risk is sufficiently minimized under our Preparedness Framework.
Unknown[1]
We conducted three distinct red team campaigns to assess gpt-5-thinking’s safety against bioweaponization applications.
Unknown[1]
This model remains on the cusp of being able to reach this capability.
Unknown[1]
Space: Let’s explore the GPT-5 Model Card