OpenAI brings its GPT models to Google Cloud, ending the one cloud one model era
OpenAI and Google Cloud announced a partnership on 21 October 2026 to put GPT models and Codex into Google Cloud's Model Garden.
OpenAI and Google Cloud announced a partnership on 21 October 2026. OpenAI's latest models and the Codex coding tool are now entering Google Cloud's Model Garden. Google Cloud customers can use GPT models without leaving the cloud environment they already work in.
1. What was announced
Three OpenAI models will be available in Model Garden: GPT-6 Astra for demanding work like coding and research, GPT-6.1 Sol which is near Astra intelligence at lower cost, and GPT-6 Luna for fast, everyday work at scale.
Purchases run through Google Cloud Marketplace. All usage appears on a single Google Cloud invoice, and companies with GCP spending commitments can draw those commitments down to pay for OpenAI models.
2. Who can use it now
The models are rolling out in private preview to selected enterprise customers. General availability is planned for the coming weeks.
Starting today, customers can also connect their local Codex CLI and Codex Desktop workflows to the Gemini platform for inference. OpenAI models will arrive in Google Antigravity and the Gemini Agent later.
3. Google Cloud becomes the meeting point for every model
With this move, Google Cloud says it is now the only hyperscale cloud provider offering all three leading closed source LLMs, alongside more than 200 other models on one platform.
For OpenAI, this completes the coverage. Its models reached AWS in April 2026 after the exclusive Microsoft cloud deal was ended. All three major clouds now carry OpenAI models. Microsoft remains OpenAI's primary cloud partner.
4. The data and security questions
Google Cloud states that customer prompts and outputs are never used to train foundation models. OpenAI models run on dedicated infrastructure behind isolated gateways, and Zero Data Retention options are coming.
OpenAI president Greg Brockman said the partnership gives customers greater choice in how they use frontier AI where they already build. Google Cloud CEO Thomas Kurian stressed having a broad selection of models for different jobs.
5. What the analysts say
Mike Leone, VP and analyst at Moor Insights & Strategy, told Axios the deal signals confidence from Google, because Gemini now has to win those customers with OpenAI sitting right next to it.
He does not see Gemini as hurt, since customers still pay Google for the cloud platform, data services and security no matter which model they pick. Once every cloud has the top models, he said, the competitive advantage goes to whoever does the best job with a company's data, governance and costs.
What it means for you
If your company runs on Google Cloud, model choice just got wider without new contracts or data migration. For everyday users the effect is indirect, but the industry direction is clear: models are no longer tied to one cloud company.
- Google Cloud users: you can try GPT models through Model Garden in private preview, general availability lands in weeks.
- Developers: the Codex CLI connection to the Gemini platform is usable today.
- IT decision makers: one invoice and GCP spend commitments across models simplify AI budgeting.
- Regular ChatGPT users: nothing changes for you directly, this is an enterprise side deal.
- More choices ahead: do not be surprised if your vendor starts asking you to pick a model per task instead of one model for everything.