1300 166 138 Tell us what you need
News / OpenAI

The models behind ChatGPT are now on Google Cloud

By Monika Tantau • Published 4 weeks ago
A woman at her desk looks at the ChatGPT home screen on a desktop monitor; the OpenAI models behind ChatGPT are now coming to Google Cloud

Key takeaways

• OpenAI and Google Cloud announced on 21 October 2026 that OpenAI models will be sold through Model Garden
• GPT 6 Astra, GPT 6.1 Sol and GPT 6 Luna are in private preview for selected enterprise customers
• Usage is billed through Google Cloud Marketplace and can draw down existing Google Cloud commitments
• OpenAI says general availability is planned in the coming weeks, and Microsoft remains its primary cloud partner

Most Read

OpenAI’s frontier models are coming to Google Cloud, but for now only a small group can use them. GPT 6 Astra, GPT 6.1 Sol and GPT 6 Luna are rolling out in private preview to selected enterprise customers in Google Cloud’s Model Garden, Google Cloud announced on 21 October. Everyone else can register interest and wait.

OpenAI says general availability is planned in the coming weeks. The deal matters because it lets companies that run on Google Cloud buy the models behind ChatGPT without signing a separate contract with OpenAI, and pay for them from cloud budgets they have already committed.

What OpenAI and Google announced

In its own post, OpenAI says Google Cloud customers will be able to buy its frontier models through Model Garden, transacted through Google Cloud Marketplace. Enterprises with Google Cloud spend commitments can draw them down for these purchases, and Google adds that usage appears on a single Google Cloud invoice.

Developers in the preview can call the models through Google’s APIs and connect local Codex workflows to Google’s platform for inference. Both companies say the models will also reach Google Antigravity, the coding environment, and Google’s Gemini business agent soon. Neither gave a date for those.

Which models are included

Google lists three models. GPT 6 Astra is designed for demanding work across coding, computer use, professional tasks and research. GPT 6.1 Sol offers close to Astra intelligence at lower cost, and GPT 6 Luna is the fast, cheaper option for everyday work at scale. Our guide to GPT 6 Astra, Sol and Luna explains which one suits which job.

Google has not published prices for OpenAI models in Model Garden. Until it does, teams should not assume the cost will match what OpenAI charges through its own API.

A phone screen titled Cloud Services showing the Amazon Web Services, Microsoft Azure and Google Cloud apps, the three clouds that now offer OpenAI models

OpenAI now spans the three big clouds

OpenAI models were already available through Microsoft Azure and Amazon Bedrock, so Google Cloud completes the set of the three largest cloud providers. OpenAI is careful to add that Microsoft remains its primary cloud partner.

For Google the win is different. It says it is now the only hyperscale cloud provider offering access to all three leading closed source model families, alongside more than 200 other models. A team can test Gemini, Claude and GPT side by side on one platform, a point our Gemini 4 Argon vs GPT 6 Astra comparison touches on.

What it means for Australian businesses

Plenty of Australian organisations already run on Google Cloud, which has regions in Sydney and Melbourne. Neither company has said whether OpenAI models will be served from those regions, so data residency is an open question. Anyone with obligations under the Privacy Act 1988 or sector rules such as APRA’s CPS 234 should ask their Google account team where processing happens before moving sensitive workloads.

Google says prompts, context and outputs stay within the customer’s data perimeter and are not used to train foundation models. Zero data retention options and confidential computing are marked as coming soon, not available yet. For regulated teams, those two features may decide whether the move makes sense.

The commercial upside is real. A business with a Google Cloud commitment can trial GPT models without a new vendor relationship, which makes it easier to choose models by workload rather than by contract.

What to watch in the coming weeks

The key signals are the general availability date, the price list in Model Garden and the regions where the models run. Watch also for the Antigravity and Gemini integrations, which would put GPT models inside tools many developers already use.

For Australian teams on Google Cloud, the practical step now is to register interest, map which workloads might benefit from a GPT model, and get answers on residency before the preview opens up.

Scroll to Top