OpenAI Adds India to API Inference Residency for GPT-6 Luna

The company says the change will let firms that must keep data in India run supported GPT-6 Luna workloads and store customer content inside the country.

Overview

OpenAI announced on Thursday that eligible Indian businesses and developers can process supported API requests and run model computation for GPT-6 Luna inside India using a new India regional endpoint.
To use the feature customers must create a new API project, select India for both storage and processing, and call the Responses or Chat Completions APIs that currently support GPT-6 Luna input and output.
OpenAI says the in‑country processing and storage target sectors with strict data‑localisation rules, such as banks and financial services, that have been reluctant to move AI into production until data stayed inside India.
The rollout is initial and model‑limited: it covers GPT-6 Luna and is available only to eligible customers, so wider enterprise impact will depend on when OpenAI expands model support, eligibility and pricing.
The move builds on OpenAI’s India push, including a Tata Group partnership to develop local data‑centre capacity, and it narrows a differentiation previously held by Indian startups that offered local hosting.