Eligible API customers can use its frontier models with Zero Data Retention, a setting designed so customer prompts and outputs are not stored by default. The company also previewed Private Safety Processing, which aims to detect misuse without compromising customer data privacy.
The announcement targets enterprise customers that want stronger models but have strict data-handling requirements. Privacy controls have become a competitive issue as companies compare AI vendors for regulated and sensitive workloads.
The important limits are eligibility and implementation detail. Zero retention is not the same as a blanket privacy guarantee for every product or user. Customers still need to confirm which models, endpoints, and safety-processing paths are covered in their own contracts.