OpenAI models are coming to Google Cloud, payable through existing Google commitments

OpenAI and Google Cloud have formed a partnership that will bring OpenAI's frontier models and its agentic coding tool Codex into Google Cloud's model catalogue, Model Garden. OpenAI announced the deal on Wednesday 21 October 2026 at 14:00 Oslo time. General availability is planned "in the coming weeks". Until then, customers can request access.

For organisations that already run their AI work on Google Cloud, this changes the calculation. They will be able to buy OpenAI models through Google Cloud Marketplace and draw the purchase down from existing spend commitments with Google. At the same time, OpenAI writes that Microsoft remains its primary cloud partner.

What the deal includes

According to OpenAI, customers will get native access to its frontier models and Codex-based agentic workflows directly in Model Garden. The integration is meant to pair the models with Google Cloud's built-in access controls, centralised data governance and infrastructure for demanding workloads.

OpenAI also says its models will "soon" be available in Google Antigravity and in Gemini Agent. That would let customers choose a model per task inside Google's own agent products. No date is given for this.

The announcement does not say which specific models are included, what they will cost on Google Cloud, or in which regions they will be available. Nor does it say anything about data processing terms or whether data can be kept in the EU or EEA.

Buying through existing cloud contracts

The most concrete point for executives is the commercial model. Purchases go through Google Cloud Marketplace. Organisations that have committed to a certain level of spend with Google Cloud can apply that commitment to OpenAI models.

In practice, this means a procurement team does not necessarily need a separate contract with OpenAI to start using the models. Budget already committed to Google can be used. For many organisations, a new vendor contract is the slowest part of adopting a new model.

Google Cloud CEO Thomas Kurian says in the announcement that the company wants to give customers a broad selection of models, "including those from OpenAI", so they can choose the right model for each job. OpenAI president Greg Brockman describes the partnership as an important step in making its models and Codex available where customers are already building.

Microsoft is still the main partner

The announcement ends with a single sentence: Microsoft remains OpenAI's primary cloud partner. OpenAI does not explain what this means for customers who currently use OpenAI models through Microsoft Azure.

Even so, the deal means OpenAI's models can be bought through more than one major cloud provider. In hogby.ai's assessment, that weakens the argument that the choice of cloud provider locks in the choice of AI model. The negotiating position of customers with several cloud contracts may also improve, but the announcement contains no prices that could confirm this.

What executives should note

• Organisations with spend commitments on Google Cloud can get access to OpenAI models without a separate OpenAI contract. Check how much of that commitment is actually unused.
• No regions are given. Anyone who must keep data in the EU or EEA should wait for documentation from Google Cloud before using the models on sensitive data.
• No pricing is given. Compare the Google Cloud offer with existing agreements with OpenAI or Microsoft before moving workloads.
• Codex on Google Cloud lets development teams use OpenAI's coding agent within Google's access controls. Agree who owns governance of such agents before they are rolled out.

Sources and media

• OpenAI: OpenAI models are coming to Google Cloud, announcement published 21 October 2026 at 14:00 Oslo time
• The source was checked on 21 October 2026.
• Illustration: AI-generated with Grok Bot / hogby.ai