GPT-6 Luna gets in-country processing in India, Japan and Australia
Regional endpoints in India, Japan and Australia now run GPT-6 Luna in country, at a 10% uplift.
OpenAI announced in-country processing for its API in India on November 12, 2026, starting with GPT-6 Luna. OpenAI's data residency guide now lists gpt-6-luna with regional processing in Australia, India and Japan, through au.api.openai.com, in.api.openai.com and jp.api.openai.com.
What changed
- In Australia, India and Japan,
gpt-6-lunasupports regional processing on Standard pay-as-you-go only. Requests must come from an API project created in that region. - Responses and Chat Completions support text and image inputs with text output. Hosted tools are unavailable for regional processing in these three regions, and image support there requires additional approval.
- India access goes through OpenAI sales. Australia and Japan have no such note in the guide.
- For requests to these endpoints, OpenAI uses Cloudflare Regional Services so TLS termination and HTTPS decryption happen in the selected region.
- System data (account data, usage statistics, billing and similar metadata) may still be stored and processed outside the region.
Price
OpenAI charges a 10% uplift on data residency endpoints for models released on or after March 5, 2026. Per million tokens, from OpenAI's pricing page on November 14, 2026:
| Endpoint | Input | Cached input | Output |
|---|---|---|---|
| Global (Standard) | $0.10 | $0.01 | $0.50 |
| Regional processing (+10%) | $0.11 | $0.011 | $0.55 |
GPT-6 Luna on RouterPlus
The RouterPlus catalog lists GPT-6 Luna as gpt-6-luna. Regional processing applies only to calls made with an OpenAI regional project key to OpenAI's regional endpoints. Per million tokens, from the catalog on November 14, 2026:
| Provider | Input | Cached input | Output |
|---|---|---|---|
| Inference.net | $0.10 | $0.01 | $0.50 |
| OpenAI (direct API) | $0.10 | $0.01 | $0.50 |
More OpenAI models are on the OpenAI lab page.
What to do
- In the OpenAI API Platform, create a new project and pick Australia, India or Japan under Data and inference residency. The region can't be changed after the project is created.
- Create an API key in that project.
- Point your client at the regional host:
import os
from openai import OpenAI
client = OpenAI(
api_key=os.environ["OPENAI_API_KEY_IN"], # key from the India project
base_url="https://in.api.openai.com/v1",
)
response = client.responses.create(model="gpt-6-luna", input="Hello")