OpenAI Models and Codex Now Available on Google Cloud
OpenAI and Google Cloud said around October 21, 2026 that OpenAI's models and its Codex coding agent are coming to Google Cloud's Model Garden, starting with a private preview for select enterprise customers. The move would put OpenAI's models on all three of the largest US cloud providers at once: AWS, Microsoft Azure and now Google Cloud, a rival that competes directly with OpenAI's ChatGPT through its own Gemini models.
⚡ Quick facts
- What's new: OpenAI's models and its Codex coding agent are rolling into Google Cloud's Model Garden in private preview
- Reported models: GPT-6 Astra, GPT-6.1 Sol and GPT-6 Luna, per multiple outlets covering the announcement — not yet independently confirmed in detail by either company
- Significance: Would put OpenAI's models on all three largest US clouds — AWS, Azure and Google Cloud
- Billing: Reportedly drawn from existing Google Cloud spending commitments via a single invoice, bought through Google Cloud Marketplace
- Timing: Private preview from around Oct 21, 2026; general availability described as "coming weeks," no firm date set
What was announced
Multiple outlets covering the announcement, including TheNextWeb, TechRadar and CryptoBriefing, reported that OpenAI is bringing its frontier models and its Codex coding agent to Google Cloud, a company that also happens to be one of OpenAI's biggest competitors through its Gemini model family. The integration runs through Google Cloud's Model Garden, the catalog Google uses to let enterprise customers deploy a mix of first-party and third-party AI models inside their own Google Cloud environment. As of this writing, neither OpenAI's nor Google Cloud's own blog carries a detailed public post laying out every specific, so this account is built from that secondary coverage rather than a single primary-source announcement.
Why it matters: OpenAI is now on all three major clouds
Before this announcement, OpenAI's models were already available through Microsoft Azure, its longtime primary cloud partner, and through AWS, which OpenAI added in April 2026 after Microsoft's period of exclusivity ended. Adding Google Cloud would complete the set of the three largest US hyperscalers. Coverage of the announcement describes Google positioning itself as the only hyperscale cloud offering access to all three leading closed-source model families in one place, since AWS and Azure customers still cannot get the complete rival lineup through either of those providers' own model catalogs.
Which models are involved — and why the names aren't fully settled yet
Secondary coverage, including reporting that cites CNBC, Axios and The Information, names GPT-6 Astra, GPT-6.1 Sol and GPT-6 Luna as the models heading into Model Garden's private preview. That said, exact naming has varied slightly between outlets, and this story is a case where the publicly available reporting is more consistent on the broad shape of the deal than on every technical detail. Treat the specific model names as reported rather than confirmed until OpenAI or Google publish their own documentation, and expect some of this to firm up once general availability arrives.
Codex, billing and what's not live yet
Alongside the models, OpenAI's Codex coding agent is reportedly part of the Google Cloud rollout, giving developers a path to Codex-based coding workflows inside Google's infrastructure. On billing, reports describe OpenAI usage on Google Cloud appearing on a single Google Cloud invoice and drawing down a customer's existing Google Cloud spending commitments, with access sold through Google Cloud Marketplace. Two features are explicitly described as not yet available: Zero Data Retention options and Confidential Compute are both listed as "coming soon" rather than live at the private-preview stage. Coverage also says OpenAI's models are slated to reach Google's Antigravity code editor and Gemini Agent, though no date has been set for either.
What this does not change
This is not OpenAI pulling back from Microsoft. Microsoft remains OpenAI's primary cloud and commercial partner, and reports say Microsoft still holds a first right of refusal on OpenAI's compute needs even though it is no longer OpenAI's sole infrastructure provider. OpenAI's cloud strategy in 2026 has increasingly been about spreading compute and distribution across multiple providers rather than consolidating around one, and Google Cloud is the latest addition to that pattern, not a replacement for AWS or Azure.
What's still unconfirmed
No exact general-availability date has been published by either company. The precise list of included models, the final shape of Zero Data Retention and Confidential Compute support, and the rollout timeline for Antigravity and Gemini Agent integration are all described in current reporting as still in motion. Readers should expect some of these specifics to change between this private-preview stage and a full public launch.
Frequently asked questions
What did OpenAI and Google Cloud actually announce?
Around October 21, 2026, OpenAI and Google Cloud said OpenAI's models and its Codex coding agent would become available to enterprise customers through Google Cloud's Model Garden, starting in a private preview with general availability expected in the coming weeks. That puts OpenAI's models on all three largest US cloud providers: AWS, Microsoft Azure and now Google Cloud.
Which OpenAI models are included?
Coverage of the announcement names GPT-6 Astra, GPT-6.1 Sol and GPT-6 Luna as the models rolling into Model Garden's private preview, alongside Codex for coding workflows. Neither OpenAI nor Google has published its own detailed model list as of this writing, so treat the exact names and version numbers as reported rather than independently confirmed, and expect them to firm up by general availability.
How will billing work?
Reports describe OpenAI usage on Google Cloud as billed through a single Google Cloud invoice, drawing down a customer's existing Google Cloud spending commitments rather than requiring a separate OpenAI billing relationship. Customers would reportedly buy access through Google Cloud Marketplace.
Does this mean OpenAI is moving away from Microsoft?
No. Microsoft remains OpenAI's primary cloud partner and reportedly still holds a first right of refusal on OpenAI's compute needs. OpenAI already added AWS as a provider in 2026 after its exclusive arrangement with Microsoft ended, and Google Cloud is the next addition to that multi-cloud approach, not a replacement for any existing partner.
What features are not yet available?
Reports on the announcement list Zero Data Retention options and Confidential Compute as coming soon to the Google Cloud integration, not available at the private-preview stage. OpenAI's models are also said to be headed to Google's Antigravity code editor and to Gemini Agent, though no firm date has been given for either.
Comments