OpenAI and Google Cloud partner to bring GPT-6 models to enterprise users

A new multi-model ecosystem

OpenAI and Google Cloud have announced a new partnership that brings OpenAI's frontier AI models directly into Google's cloud infrastructure. This agreement allows enterprise customers to access the GPT-6 family of models alongside Google's own tools. The move is designed to give organizations more flexibility in choosing the right AI for specific workloads.

The initial rollout includes three distinct models tailored for different needs. GPT-6 Astra is built for high-demand tasks like scientific research and professional coding. GPT-6.1 Sol provides similar intelligence at a lower price point for agentic coding, while GPT-6 Luna is a faster, cheaper option for high-volume everyday tasks.

By making these models available in the Google Cloud Model Garden, the platform allows users to switch between models without signing new vendor contracts. This setup helps businesses balance performance, response speed, and cost. As EdTech Innovation Hub reported, the integration aims to provide a broad selection of models for any given job.

Simplified billing and access

One of the primary benefits for large organizations is the streamlined commercial arrangement. Customers can purchase OpenAI model usage through the Google Cloud Marketplace. This means all charges will appear on a single Google Cloud invoice rather than requiring separate billing systems for different providers.

The partnership also allows companies to use their existing spending commitments more effectively. Brian Kane, Director of Cloud Partnerships at OpenAI, noted that customers can draw down spend commitments with both OpenAI and Google Cloud simultaneously. This helps enterprises maximize the value of their pre-negotiated cloud contracts.

Google Cloud is positioning this as a way to provide a unified enterprise AI platform. In addition to the OpenAI models, the platform offers more than 200 first-party, third-party, and open models. This variety allows developers to experiment with different technologies while staying within a single ecosystem.

Security and data governance

For many enterprises, the biggest hurdle to using AI is keeping sensitive information private. Google is addressing this by incorporating its existing security and data governance infrastructure into the OpenAI offering. The company states that customer prompts, context, and outputs will remain within the designated Google Cloud data perimeter.

To ensure privacy, queries will run through isolated gateways on dedicated Google Cloud infrastructure. This setup prevents the data from being used to train any foundation models. The platform also provides authenticated identities, centralized access controls, and audit logging for all interactions between models and agents.

While some features are still being finalized, Google is working on additional safeguards. These include Zero Data Retention options and Confidential Compute support. These tools are intended to provide extra layers of protection for organizations handling highly sensitive information.

Developer tools and integrations

The partnership extends beyond just providing raw models to the cloud. It includes integrations with Google Antigravity, which is an AI coding platform. Developers can also connect local Codex CLI and Desktop workflows directly to Google's Gemini platform for inference during the private preview.

Google plans to expand these integrations further into Gemini Agent workflows. These tools are designed to help with AI-assisted software development and enterprise agent tasks. By connecting these systems, developers can use models from different providers within a single development environment.

This multi-provider approach is a key part of Google's strategy to offer a broad selection of models. Thomas Kurian, CEO of Google Cloud, stated that the goal is to help customers choose the right model to deliver optimal quality and lower costs. The integration makes it easier for teams to build complex applications that might require different models for different steps.

Rollout and next steps

The GPT-6 models are currently entering a private preview for selected enterprise customers. This means the service is not yet available to the general public. Google Cloud is currently accepting expressions of interest, and OpenAI has opened a separate request process for access.

Existing customers can also reach out to their specific account managers at either OpenAI or Google Cloud to request access. The companies expect to move to general availability in the coming weeks. However, no specific dates have been confirmed for the wider release of these models.

Following the initial release, Google plans to roll out the specific integrations for Google Antigravity and Gemini Agent. These updates will follow the general availability of the GPT-6 models. The timeline for these additional features remains unconfirmed as the companies work on the final integrations.