OpenAI's models are now on sale inside Google Cloud. On October 21, Google Cloud announced that GPT-6 Astra, GPT-6.1 Sol and GPT-6 Luna are rolling out in its Model Garden, and OpenAI confirmed the partnership the same day. For now it is a private preview for select enterprise customers. OpenAI says general availability is planned "in the coming weeks."
The deal is unusual because the two companies compete directly. Google sells Gemini. OpenAI sells GPT models and ChatGPT. Now Google will resell its rival's models next to its own, and OpenAI gets another route into large companies that already run on Google's cloud.
What is available today
Google lists three OpenAI models in Model Garden, its catalog of first-party, third-party and open models:
- GPT-6 Astra: Google describes it as built for demanding coding, computer use, professional tasks and scientific research.
- GPT-6.1 Sol: described as "near-Astra intelligence at lower cost" for agentic coding, computer use and professional work.
- GPT-6 Luna: a fast, cost-efficient model for everyday work at scale.
Google says customers in the preview can build with these models through its APIs starting today. It also says developers can connect their local Codex CLI and Codex Desktop workflows to Google's Gemini platform for inference. That means a team that already uses OpenAI's coding agent can point it at Google Cloud instead of OpenAI's own API. Access goes through an interest form, so most teams cannot just switch it on yet.
Neither company published Google Cloud prices for the models in its announcement.
What is coming soon
Several pieces are announced but not live:
- General availability, which OpenAI says is planned in the coming weeks.
- Google Antigravity, Google's coding environment, will get OpenAI models.
- Gemini Agent will offer OpenAI models as a choice, according to OpenAI. Google's post says the models will come to "Gemini for business."
- Zero Data Retention options are listed by Google as "coming soon."
- Confidential Compute for the OpenAI models is also listed as "coming soon."
How billing works
This is the part most enterprise buyers will care about. OpenAI models are sold through Google Cloud Marketplace, so usage shows up on the same Google Cloud invoice as everything else. Google says the spend draws down existing Google Cloud commitments and counts toward its Flexible Savings Plans across model providers.
For a company that has already signed a large multi-year cloud commitment with Google, that matters. It can use OpenAI models without a separate contract with OpenAI and without new money outside the budget it has already committed.
Data and security claims
Google makes several specific promises about how OpenAI workloads run:
- Prompts, context and outputs stay inside the customer's data perimeter and are not used to train foundation models.
- The OpenAI models run on dedicated Google Cloud infrastructure, and queries go through isolated gateways so that data stays inside Google Cloud.
- Model calls run under an authenticated identity, with VPC Service Controls and unified audit logging for administrators.
These are Google's descriptions. Regulated customers will want to confirm the details in their contracts, especially since the Zero Data Retention option is not available yet.
Why it matters
For Google, the announcement lets it say it is, in its words, "the only hyperscale cloud provider offering access to all three leading closed source LLMs." Google did not name the three. Google Cloud CEO Thomas Kurian framed it as choice: "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, helping to deliver optimal quality and lower costs."
If a customer prefers GPT models for a task, that spend can now still run through Google Cloud. CNBC reports that Google's cloud revenue growth accelerated to about 82% in the second quarter, ahead of Amazon at 37% and Microsoft at 42%.
For OpenAI, this completes its spread across the three big US clouds. CNBC notes that Amazon began hosting OpenAI models in April, after Microsoft's exclusive license on OpenAI's IP ended. OpenAI's post ends with a reminder that "Microsoft remains OpenAI's primary cloud partner." OpenAI president Greg Brockman said: "This partnership is an important step in making our most capable models and Codex available where customers are already building, giving them even more choice in how they use frontier AI as we grow the agentic ecosystem together."
For developers and IT teams, the practical change is less about model quality and more about procurement. If your company is a Google Cloud shop, you may soon be able to test GPT-6 Astra against Gemini 4 Argon in the same console, under the same identity and audit controls, on one bill.
What to watch
- When general availability lands, and whether Google publishes per-token prices that match or differ from OpenAI's own API.
- Whether the newest OpenAI models arrive on Google Cloud at the same time as on OpenAI's platform. Google says "more coming soon" but gives no schedule.
- How quickly the OpenAI options appear inside Antigravity and Gemini Agent, where they would sit next to Google's own models.
If you want in now, Google's post links an interest form for the private preview.
Related on llms.blog: our Gemini 4 Argon explainer covers the Google model that now sits beside these OpenAI models in Model Garden, and our report on Emirates NBD running OpenAI models on in-country UAE inference shows another route OpenAI is taking beyond Microsoft.
