OpenAI puts its frontier models and Codex on Google Cloud's Model Garden
OpenAI and Google Cloud announced a partnership that will bring OpenAI's frontier models and its Codex agentic workflows into Google Cloud's Model Garden. The models will be available natively in the platform, backed by Google Cloud's data governance and high-performance infrastructure, and customers will transact through Google Cloud Marketplace, according to OpenAI's announcement of October 21.
How the purchase works
Model Garden is the catalog of first-party and third-party models that developers deploy on Google Cloud, so the placement puts OpenAI's frontier line in front of the teams already building on that environment. OpenAI frames it as giving those teams access to its models where they already work.
The billing mechanics matter more than the catalog slot. Google Cloud customers will be able to draw down existing GCP spend commitments to pay for OpenAI models. In practice, a company's frontier inference from OpenAI can run on the same committed-spend framework as the rest of its cloud bill, with no separate vendor relationship to stand up.
General availability is planned in the coming weeks, and OpenAI has opened an access request in the meantime. Microsoft remains OpenAI's primary cloud partner, and the announcement says the relationship is unchanged.
Governance for regulated industries
For organizations in regulated industries, the pitch is control. The integration pairs OpenAI's frontier models with Google Cloud's built-in access controls and centralized data governance, and OpenAI says the setup is meant to support demanding enterprise workloads. The company's argument is that this is what lets enterprises put frontier models on sensitive data, with control over who can access it and how it is used, rather than keeping those models out.
OpenAI also lists the workloads the access is meant to unlock: modernizing the software development lifecycle, building agentic workflows at scale, and Codex-driven development. The models will land beyond the Model Garden as well. OpenAI says they will soon be available in Google Antigravity and the Gemini Agent, which would give Google's enterprise customers a model choice per task instead of a single default.
A second distribution route
This is not the first cloud commitment OpenAI has plugged into another vendor's. Its own newsroom links to a June post about accessing OpenAI models and Codex through Oracle cloud commitments. Alongside Microsoft's first-party position, the pattern is a lab spreading its inference across the clouds where its customers' budgets already sit, with a stated path from evaluation to broader adoption.
Greg Brockman, OpenAI's president and co-founder, called it an important step in making "our most capable models and Codex available where customers are already building." Thomas Kurian, Google Cloud's CEO, said the company is focused on giving customers a broad selection of models, "including those from OpenAI," to pick the right model for any given job.
What to take from this
If you are deciding where to run frontier inference, the option set just widened. A team with an existing GCP commitment can evaluate OpenAI models inside the same governance and billing stack it already runs, and Google's agent products will let you route tasks across models. What the announcement does not publish is pricing or capacity limits, and neither company has given a date beyond "coming weeks." The practical move for now is to file the access request and track general availability.
Source: openai.com
More news
Have something in mind?
Start a conversation