Technology · NEWS EXPLAINER
OpenAI models are coming to Google Cloud: what Model Garden access changes for enterprises
GPT-6 Astra, GPT-6.1 Sol and GPT-6 Luna are entering private preview on Google Cloud, with Marketplace purchasing and broader availability planned in the coming weeks.
OpenAI and Google Cloud are opening a new enterprise distribution route for OpenAI frontier models and Codex workflows through Google Cloud’s Model Garden.
The first models are rolling out in private preview rather than universal access. The practical change is procurement and governance: eligible customers can evaluate the models inside Google Cloud and, when generally available, buy them through Cloud Marketplace using existing GCP commitments.
What was announced on October 21
OpenAI and Google Cloud announced that OpenAI frontier models and Codex agentic workflows will be offered through Google Cloud’s Model Garden. Reuters independently reported the agreement at 12:01 UTC. It is a distribution and enterprise-access partnership, not an acquisition or an announcement that every Google Cloud account can use the models immediately.
Private preview comes before broad access
Google Cloud says GPT-6 Astra, GPT-6.1 Sol and GPT-6 Luna are rolling out now in private preview to selected enterprise customers. OpenAI says general availability is planned in the coming weeks. Because no exact general-availability date was supplied, organizations should treat current access as controlled preview availability.
Procurement may be the biggest operational change
When the offering reaches Marketplace, customers will be able to transact through Google Cloud and draw down eligible GCP spending commitments. That can reduce a procurement barrier for enterprises already governed and budgeted around Google Cloud, but it does not disclose model pricing, quotas, regions or service-level terms.
What developers can do first
Google Cloud says selected customers can build with the preview models through its APIs and connect local Codex CLI and Desktop workflows to the Gemini platform for inference. OpenAI also says its models will later appear in Google Antigravity and Gemini Agent, giving teams another route to select models for coding and agent workflows.
Why governance is central to the pitch
Both companies emphasize access control, centralized governance and high-performance infrastructure, especially for regulated industries. Those platform controls can simplify identity and policy management, but each customer still needs to verify data handling, residency, retention and sector-specific obligations for its exact deployment.
What the deal does not change
OpenAI explicitly says Microsoft remains its primary cloud partner. Reuters describes Google as the latest major hyperscaler to distribute OpenAI models after Amazon Web Services. The new route expands channel choice; it does not make Google the exclusive or primary partner, and it does not replace direct OpenAI or other cloud options.
