OpenAI Models on Google Cloud: Three GPT-6 Generation Models, Codex Integration and Availability

Screenshot of the official Google Cloud — Google Cloud expands model choice and flexibility, bringing OpenAI models to customers page — Google Cloud — Google Cloud expands model choice and flexibility, bringing OpenAI models to customers · SourceCaptured: October 26, 2026

Screenshots show the official pages on the capture date and may differ from the pages at publication. Select an image to enlarge it.

On October 21, 2026, OpenAI and Google Cloud announced a partnership to offer OpenAI's models on Google Cloud. Three models, GPT-6 Astra, GPT-6.1 Sol and GPT-6 Luna, are being added to Model Garden, the place in Google Cloud where customers choose and use AI models. The rollout begins as a limited preview for select businesses, and OpenAI says it plans to start general availability within the next few weeks.

What was announced

According to OpenAI's announcement, the company is bringing its frontier models and workflows built on its coding agent Codex to Google Cloud's Model Garden. Businesses on Google Cloud will be able to use OpenAI's models without leaving the development environment they already work in.

Google Cloud says this makes it the only major cloud provider to offer all three leading closed models (models whose weights are not published). Google Cloud already offers its own Gemini and Anthropic's Claude, and OpenAI's models now join them. Model Garden gives access to more than 200 models in total, spanning Google's own, third-party and open models.

The models on offer

The three new additions belong to OpenAI's GPT-6 generation. Google Cloud describes them as follows.

GPT-6 Astra: for demanding work such as coding, computer use, professional tasks and scientific research
GPT-6.1 Sol: intelligence close to Astra's at a lower price
GPT-6 Luna: a fast, cost-efficient model for high-volume everyday work

Pricing on Google Cloud was not given in the announcement. On OpenAI's API, Sol costs one-fifth as much as Astra (see also our article on GPT-6.1 Sol Ultrafast).

Codex and other Google services

According to Google Cloud, businesses in the limited preview can use these models through the API starting on the day of the announcement. They can also connect the Codex CLI (the command-line version) or desktop app on their own machines to Gemini platform, Google Cloud's enterprise AI platform, and use it for inference (the process of having a model generate answers).

Google also says OpenAI's models will soon be available in Google Antigravity, its AI coding environment, and in Gemini for business, its enterprise offering (referred to as "Gemini agents" in OpenAI's announcement). Users will be able to pick between OpenAI and Google models task by task.

Purchasing and data handling

OpenAI's models are purchased through Google Cloud Marketplace. Charges appear on the customer's Google Cloud bill and also count toward the spending commitments businesses have made with Google Cloud.

On data handling, Google Cloud highlights the following points.

Prompts, documents and outputs stay within the data boundary the customer specifies and are not used to train foundation models
OpenAI's models run on dedicated Google Cloud infrastructure
Zero data retention and Confidential Compute, which encrypts data even while it is being processed, are coming soon

Greg Brockman, OpenAI's president and co-founder, said: "This partnership is an important step in making our most capable models and Codex available where customers are already building." Thomas Kurian, CEO of Google Cloud, said: "We are focused on giving customers a broad selection of models, including those from OpenAI, so they can choose the right model for any given job," adding that this helps them optimize quality and reduce costs.

The relationship with Microsoft

OpenAI's models have long been offered on Microsoft's cloud, Azure. In this announcement, OpenAI states plainly that "Microsoft remains OpenAI's primary cloud partner."

At the same time, OpenAI has been expanding where its models are available. In June, it made its models and Codex available on Oracle Cloud, payable through businesses' existing spending commitments. According to CNBC and The Next Web, Amazon's cloud, AWS, also began offering OpenAI's models in April, after Microsoft's exclusive arrangement ended. With Google Cloud, OpenAI's models are now available on all three of the major US clouds.

CoAI's take

Companies that build AI models and companies that run clouds are also competitors. Google, which has Gemini, selling OpenAI's models on its own cloud is a sign that the era when a company's choice of cloud dictated which models it could use is coming to an end.

For businesses in Japan that already use Google Cloud, the advantage is being able to try OpenAI's latest GPT models without adding new contracts or billing arrangements. For now, however, this is a limited preview for select businesses, and there is no word yet on when it will be available from Japan, in which data center regions, or at what price. Zero data retention and similar features are also still "coming soon." It makes sense to check whether the conditions your internal policies require will be met, and to wait for general availability before evaluating it seriously.