Data Residency for the OpenAI API
This FAQ covers common questions about OpenAI's new Data Residency feature. Eligible customers can now opt to store and process data with the United States or within Europe.
For more information about data residency for ChatGPT, please refer to our other help center article instead.
Eligibility
Which customers can enable European Data Residency when using the API?
OpenAI customers that have an Enterprise agreement and that are eligible for Zero Data Retention can enable European Data Residency.
How does European Data Residency work?
Your requests and data for eligible endpoints and models will always be processed by OpenAI services running in the European Economic Area and Switzerland. If your request is to an Endpoint that requires data persistence to function, such as /v1/batches, then that data is always persisted in data centers within Europe.
Some non-sensitive data, such as billing or usage analytics, will be processed and stored outside Europe.
Which models and features are eligible for European Data Residency?
The following models and API services are eligible for European Data Residency today.
Supported Services | Supported Model Snapshots | Region | Region |
|
| United States | Europe (EEA + Switzerland)* |
/v1/chat/completions | o3-mini-2025-01-31 o1-2024-12-17 o1-mini-2024-09-12 o1-preview gpt-4o-2024-11-20 gpt-4o-2024-08-06 gpt-4o-mini-2024-07-18 gpt-4-turbo-2024-04-09 gpt-4-0613 gpt-3.5-turbo-0125 | Yes | Yes
|
/v1/images/generations
| dall-e-3 | Yes | Yes |
/v1/audio/transcriptions /v1/audio/translations /v1/audio/speech | tts-1 whisper-1 | Yes | Yes |
/v1/moderations | Text-moderation-007 omni-moderation-latest | Yes | Yes |
/v1/batches | o3-mini-2025-01-31 o1-2024-12-17 o1-mini-2024-09-12 o1-preview gpt-4o-2024-11-20 gpt-4o-2024-08-06 gpt-4o-mini-2024-07-18 gpt-4-turbo-2024-04-09 gpt-4-0613 gpt-3.5-turbo-0125 | Yes | Yes |
/v1/embeddings | text-embedding-3-small text-embedding-3-large text-embedding-ada-002 | Yes | Yes |
/v1/fine_tuning/jobs /v1/files | gpt-4o-2024-08-06 gpt-4o-mini-2024-07-18 | Yes | Yes |
Scale Tier |
| Yes | Yes |
Structured Outputs (excluding schema) |
| Yes | Yes |
Supported Input Modalities |
| Text Image Audio/Voice | Text Image** Audio/Voice |
* EU Region support is contingent on Customer being approved for Zero Data Retention for the applicable endpoints
** Zero Data Retention for image inputs is not included in standard Zero Data Retention and requires approval for enhanced Zero Data Retention.
In which countries does inference physically occur?
Inference is restricted to the EEA (European Economic Area) plus Switzerland.
Will European Data Residency be available without ZDR?
We're currently exploring the possibility of providing European Data Residency without requiring ZDR. If you have feedback or a strong need for this capability, we’d love to hear from you—feel free to contact our team.
How do I get started?
To get started, please reach out to your account representative at OpenAI, or get in touch with us here.
Functionality
How do I set up European Data Residency?
Create a new project and select EU as the geography. Resources, such as inference engines or file storage under that Project, reside in Europe. You cannot migrate an existing Project between regions.
Next, use https://eu.api.openai.com as your base URL instead of the default endpoint. All prompts and completions flow only through EU-based GPU infrastructure, and file storage (for features like Batch API) is in Azure Blob Storage geographically located within Europe. Metadata (e.g., billing, organization details) is still stored in the US.
What happens if I send traffic to the wrong endpoint?
If you send requests from an EU-located Project’s credentials to the US endpoint, you'll see an error. Ensure you use the correct endpoint (api.openai.com for US or eu.api.openai.com for Europe).
Can I use both US and Europe regions simultaneously?
Yes. Your Organization can contain Projects in the US and Projects in Europe at the same time. Traffic for each Project must route to the corresponding endpoint.
Will there be region-specific differences in features, models, or rate limits?
We strive for parity between our US and European regions. However, some capabilities—such as alpha and beta releases—may become available in the US first. Over time, we plan to align features as closely as possible.
Does Europe Data Residency reduce latency for customers operating in Europe?
Yes, Europe-based users should experience a noticeable reduction in request latency compared to using US-based endpoints. This improvement is designed to bring response times more in line with those of US-based customers through a reduction in geographical latency.
Does authentication differ between regions?
You will have separate API keys for each Projects created in Europe. The keys for a US-region Project won’t work on the Europe endpoint, and vice versa.
Are there plans for further regional expansions outside Europe?
We're currently exploring the possibility of expanding to additional regions. If you have feedback or a strong need for expansion in a specific region, we’d love to hear from you - feel free to contact our team!
Pricing
Are there any pricing differences when European Data Residency is enabled?
No. Your usage will be priced similarly across the US and Europe.
Is Scale Tier supported in the Europe region?
Yes. Scale Tier is fully supported. Rate limits in the Europe region match those in the US.