OpenAI sets a November 20 approval requirement for model-access intermediaries
OpenAI’s updated usage policies require prior review and approval for services that give third parties direct model access from November 20, while excluding applications merely because they use OpenAI models to power features.
OpenAI’s October 21 policy announcement sets a November 20, 2026 effective date for prior review and approval of model-access intermediaries, including model routers and AI gateways.
The operative policy covers services enabling third parties to access OpenAI models directly. It expressly says an application does not fall under that requirement solely because it integrates OpenAI models into its features or functionality. [1][2]
The change makes distribution arrangements part of infrastructure planning. OpenAI says partners must maintain consistent safeguards, investigate violations and coordinate enforcement across services. A technically interchangeable API endpoint therefore may still have provider-specific approval obligations. The policy describes OpenAI’s conditions for its services; it is not a general legal licensing requirement for all model gateways. [1][2]
Teams operating a direct-access service should compare their service’s function with the published scope and confirm the required approval through their existing provider relationship before the effective date. Teams buying through a gateway can ask whether the intermediary’s approval covers that access. The announcement does not establish that any named third-party gateway has been approved, or give a public approval turnaround guarantee. [1][2]
Take it further
Check whether your service is a model-access intermediary For direct-access gateway operators and buyers. Effective November20 2026; source supplies no time-of-day or timezone. Confirm scope and approval through the provider relationship.
Sources & details 3 sources
Discuss in your voice app
Take the reporting, visual explanations and original sources into a conversation. Copy the brief, paste it into your preferred assistant, then start its voice mode.
Paste the brief into your voice app to discuss this reporting.Reporting sources
- OpenAI · access and safety policy announcement
October21. States November20 effective date and explains routers/gateways and coordinated safeguards.
- OpenAI · updated usage policies
Operative text displays effectiveNovember20 2026 and explicit application-feature exception. No timezone or approval SLA stated; provider policy, not legislation.
- OpenAI Help · effective-date notice
Separate official notice confirmsNovember20 2026 effective date and that previous policies remain effective until then. Same policy issuer, not external legal interpretation.