Broad access through consistent standards

Effective November 20, 2026, services offering direct access like model routers or AI gateways must obtain OpenAI review and approval.
Key points
- Hosting and reseller partners must maintain safeguards consistent with OpenAI's own and coordinate cross-service enforcement
- OpenAI is introducing SafetyIDs to let developers identify specific end users during misuse investigations without sharing their identities
Why it matters
AI gateway operators must get formal approval from OpenAI, while developers gain an anonymized mechanism to prevent entire accounts from being banned due to individual abusive users.
Sources
- Broad access through consistent standards (OpenAI, Oct 21, 2026, 10:07 PM)