OpenAI's models are coming to Google Cloud, and a first batch is already rolling out. Google's announcement, dated Oct. 22, says GPT-6 Astra, GPT-6.1 Sol and GPT-6 Luna are rolling out now in private preview to select enterprise customers in Model Garden, with more coming soon.

What is live and what is coming

  • Live now: private preview of the three models in Model Garden, for select enterprise customers (Google).
  • Coming: more coming soon, plus availability in Gemini for business and Google Antigravity (Google). OpenAI's announcement also names the Gemini Agent.
  • Not published: prices and regions.

CNBC adds that OpenAI's models launched on AWS in April.

The three models

  • GPT-6 Astra: the top model, pitched for demanding coding, computer use, professional and scientific work.
  • GPT-6.1 Sol: the cheaper option, with 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.

These are Google's descriptions, and no prices were published.

How enterprises buy them

Google says the models are sold directly through Google Cloud Marketplace, appear on a single Google Cloud invoice, draw down existing cloud commitments and count toward Google's Flexible Savings Plans, its discount program, across model providers.

Customers can build through Google's APIs and connect local Codex CLI and Desktop workflows to the Gemini platform for inference. Interested companies can use Google's interest form.

OpenAI's post says access is backed by Google Cloud's data governance, and built-in access controls, and aims it at regulated industries. It links the full OpenAI announcement, which adds Codex agentic workflows to the offer and says general availability is planned in the coming weeks.

Why it matters

For buyers, the practical change is procurement. Once a company has preview access, it can try OpenAI models without a separate vendor contract, and the spend draws down its existing Google Cloud commitment. The catch: with no published prices or regions, a team can't budget or check data-residency needs until it is in the preview.

For OpenAI, the deal means it no longer depends on one cloud to reach enterprise buyers. OpenAI's own announcement still says "Microsoft remains OpenAI's primary cloud partner," so this reads as OpenAI going multi-cloud, not leaving Microsoft. Google claims it is now the only large cloud provider offering all three leading closed-source LLMs, alongside more than 200 first-party, third-party and open models. That is Google's own claim.

For related coverage, see OpenAI's coding agents weakening their own safety monitoring and Google's Gemini Agent for enterprise Workspace.