OpenAI models are reportedly expected to become available through Google Cloud Model Garden, with GPT models also expected in Google’s Antigravity coding tool. Google and OpenAI have not yet confirmed the timing, model lineup, pricing, regions, or deployment details.
Key takeaways
- - Ars Technica reports that OpenAI models may be offered through Google Cloud Model Garden and Google Cloud Marketplace.
- - GPT models are also reportedly planned for Google’s Antigravity coding tool, but the supported versions and capabilities are unconfirmed.
- - The reported expansion could give Google Cloud customers another model provider alongside offerings such as Gemini, Claude, Llama, Mistral, and Gemma.
- - Model names, launch timing, pricing, regions, service limits, and data-handling terms should be treated as unconfirmed until official details are published.
- - Customers should verify billing, security, deployment, and workload-routing requirements before planning production use.
OpenAI models are expected to become available through Google Cloud Model Garden in the coming weeks, according to reporting by Ars Technica. The report also says GPT models are expected to appear in Google’s Antigravity coding tool. If the plans proceed, customers will gain another way to use OpenAI models alongside Google’s and other providers’ offerings - without relying on a single AI model family or cloud platform.
The details, however, are not all settled. Customers should treat the timing, exact model lineup, pricing and deployment options as pending confirmation until Google or OpenAI publishes official information.
What OpenAI’s arrival in Google Cloud Model Garden is expected to include
Model Garden is Google Cloud’s catalog for discovering and deploying AI models. Ars reports that OpenAI models are expected to be hosted on Google infrastructure and purchased through Google Cloud Marketplace. That could make access more convenient for organizations already managing cloud services and spending through Google.
The report also notes that the arrangement may be relevant to customers with existing Google Cloud spending commitments. Whether purchases of these models count toward a particular commitment, and under what terms, remains something customers will need to verify with Google.
For teams considering OpenAI models on Google Cloud, the practical draw is a potential route to evaluate or deploy them within an existing cloud environment. But the report does not specify operational details such as supported regions, service limits, data-handling terms, or how the models will be configured and monitored. Those details matter for production workloads, not just initial access.
Where Antigravity fits into Google’s OpenAI plans
The report says GPT models are also expected to be offered in Google Antigravity, Google’s agent-powered coding tool. That would give developers another potential model option within a tool designed to support coding work.
The report does not establish a release date, which GPT versions will be included, or what capabilities and controls will be available in Antigravity. It is therefore too early to assume that the tool will support a particular workflow, or that its model choices will match those offered through Model Garden.
How OpenAI adds to Google Cloud’s multi-model lineup
Google Cloud already offers a range of models through Model Garden. The reported lineup includes Google’s Gemini models, as well as options from Mistral, Meta’s Llama, Google’s Gemma, and Anthropic’s Claude. Google is quoted as saying it is the only large cloud provider offering access to all three leading closed large language models.
Adding OpenAI would broaden that range and could help organizations compare models or select different ones for different tasks. A company might, for example, use one model for code assistance and another for summarizing internal documents, subject to the models’ capabilities and the organization’s security requirements. Access to multiple providers does not, by itself, guarantee that workloads can be switched seamlessly; integration, governance, cost and performance still need evaluation.
There is also uncertainty around the reported model names. Ars describes a full suite of GPT 6.1 models called Astra, Sol and Luna, while a forum commenter disputes that lineup and says Astra and Luna are GPT 6 models, not GPT 6.1. These names and versions should be treated as unverified until an official announcement confirms them.
What cloud customers still need to confirm
Before planning a migration or committing workloads, customers should check for confirmed answers on:
- Launch timing and supported model versions
- Pricing, billing terms and whether purchases apply to existing cloud commitments
- Regional availability and service limits
- Deployment, security and data-handling details
- How workloads can be routed among OpenAI, Gemini, Claude and other models
For teams evaluating OpenAI models on Google Cloud, check Google Cloud’s official announcements before making plans. Confirmed launch timing, model versions, pricing, regional availability and deployment details will determine whether the offering fits a particular workload.
By the numbers
Google says it is the only large cloud provider offering access to all three leading closed large language models.
This is a Google statement quoted in the article; it describes Google’s current multi-model positioning and is not independent comparative verification.
The reported OpenAI expansion covers two Google products: Cloud Model Garden and the Antigravity coding tool.
Ars Technica is the source of the reported plans. Neither product’s OpenAI availability, model versions, or launch timing is confirmed in the article.
Step by step
- 01
Monitor official announcements
Check Google Cloud and OpenAI announcements for confirmation of launch timing, supported model versions, and availability in Model Garden or Antigravity.
- 02
Verify billing and commitment terms
Ask Google whether purchases will be billed through Google Cloud Marketplace and whether they qualify toward your existing cloud spending commitment.
- 03
Review deployment and data controls
Confirm supported regions, service limits, security controls, data-handling terms, and monitoring options before testing sensitive or production workloads.
- 04
Evaluate models against your workload
Compare confirmed models for task quality, cost, performance, and governance needs, and test how your systems would route work among providers.
Frequently asked questions
When will OpenAI models be available on Google Cloud?
There is no officially confirmed launch date for OpenAI models on Google Cloud. Ars Technica reports that availability through Model Garden is expected in the coming weeks, but customers should wait for Google or OpenAI to confirm timing.
Where are OpenAI models expected to appear in Google’s products?
OpenAI models are reportedly expected in Google Cloud Model Garden, and GPT models are also reportedly planned for Antigravity. The reported Model Garden route may involve Google Cloud Marketplace, but product availability and purchasing details remain unconfirmed.
Which GPT models will Google Cloud offer?
The GPT models expected to be offered have not been officially confirmed. Ars Technica reports names for a model lineup, but the article notes that those names and versions are disputed and should be treated as unverified.
Will OpenAI model purchases count toward Google Cloud commitments?
It is not yet confirmed whether OpenAI model purchases will count toward existing Google Cloud spending commitments. Customers should verify eligibility and billing terms directly with Google, since they may depend on the final offering and customer agreement.
