OpenAI has made local processing available for its API in India, starting with the GPT-6 Luna model, Fonearena reported on 12 November 2026. Eligible enterprises and developers can now run supported workloads inside the country. OpenAI local processing covers both where data is stored and where each answer is computed.

For Indian businesses, this removes a common blocker to using AI at scale. Banks, insurers and other regulated firms often need data to stay inside India. Until now, residency mainly covered stored content, while the model could still run elsewhere.

Quick answer

  • What changed: API requests can now be received and answered on infrastructure in India.
  • First model: GPT-6 Luna is the only model named so far.
  • Who can use it: eligible enterprises and developers, with criteria not spelled out.
  • How to switch on: create a new API project and pick India as the region.
  • Why it matters: regulated sectors such as banking and insurance get a clearer path to production.

What changes for developers

Earlier, data residency in India kept supported customer content stored in the country. Now local coverage reaches the full inference process. API requests arrive on domestic infrastructure, and the computation that writes the response runs there too. Supported customer content also stays stored in India.

Setup is simple. A developer creates a new API project and selects India as the region for both storage and runtime processing. Traffic then goes through a regional India endpoint. Both the Responses API and the Chat Completions API work with it.

Why OpenAI local processing matters for regulated firms

Data location often decides whether an AI pilot reaches production. Banking, financial services and insurance firms face the strictest checks. The report says the launch targets these compliance, regulatory and technical needs. However, it does not name a specific regulation.

Nitin Bawankule, OpenAI’s Head of Enterprise for India, said local inference gives firms more control over where supported workloads run. That control is the real product here. A model that answers well but runs abroad can still fail a compliance review.

Use cases and open questions

The report lists document analysis, automated drafting and data synthesis on internal workflows. These tasks often involve contracts, customer records or internal reports. Teams can now run them without sending payloads to overseas data centres.

Some details are still unclear. The article does not define which enterprises are eligible, or which workloads count as supported. Other models are not named either. Readers planning a rollout should check OpenAI’s own documentation first.

Where this fits in OpenAI’s India push

The launch is part of OpenAI for India, a programme that includes work with the Tata Group on domestic data centre capacity. The company says India joins Japan and Australia, which recently received similar regional infrastructure. It also sits alongside residency features offered through AWS Bedrock.

The takeaway is practical. Compliant AI projects depend as much on where data runs as on which model answers. So OpenAI local processing may matter more to Indian IT and finance teams than any single feature of the model itself.

FAQ

What is OpenAI local processing in India?

It means supported API requests are received and answered on infrastructure inside India. Supported customer content also stays stored in the country. This goes beyond earlier data residency, which covered storage.

Which OpenAI model supports it?

GPT-6 Luna is the first model in the rollout. The report does not name any other model yet.

How do I use local processing in India?

Create a new API project and select India as the region for storage and runtime processing. Requests then route through the regional India endpoint. Your account must be eligible, and OpenAI has not published the criteria in the report.