Google Cloud opens a private preview of OpenAI's GPT-6 models

The Google Cloud post is bylined Michael Gerstenhaber, VP of product management for Gemini[2]. It lists three OpenAI models:

• GPT-6 Astra, "designed for demanding work across coding, computer use, professional tasks, and scientific research."
• GPT-6.1 Sol, "near-Astra intelligence at lower cost across agentic coding, computer use, and professional work."
• GPT-6 Luna, "a fast, cost-efficient model for everyday work at scale."

Those descriptions are Google's[2]. OpenAI's post speaks only of "frontier models" and Codex, and of "our most capable models"[1].

Google also writes that it is "now the only hyperscale cloud provider offering access to all three leading closed source LLMs, alongside more than 200 first-party, third-party, and open models on a unified enterprise AI platform"[2]. The sentence does not name the other two closed models.

Google says the three models "are rolling out now in private preview to select enterprise customers" in Model Garden, "with more coming soon"[2]. "Starting today," it says, customers can build with the models through Google Cloud APIs and "connect their local Codex CLI and Desktop workflows directly to our Gemini platform for inference." The post points to a form for customers who want access[2].

The same post says that, coming soon, the models will be available in Gemini for business and Google Antigravity[2].

OpenAI's wording is parallel and not identical. It says the partnership brings frontier models and "Codex agentic workflows" into Model Garden, and that OpenAI models "will soon be available in Google Antigravity and the Gemini Agent"[1]. It uses "the Gemini Agent," where Google's post says "Gemini for business." It says general availability is planned in the coming weeks, and it points readers to a request-access link[1]. It does not say the Codex CLI connection is live on 21 October. That sentence is in Google's post[2].

OpenAI says customers will purchase the models via Model Garden, "transacted through the Google Cloud Marketplace," and that enterprises with Google Cloud spend commitments will be able to draw those commitments down for these purchases[1].

Google says the same Marketplace path, and adds three billing details: usage appears on a single Google Cloud invoice, it draws down existing cloud commitments, and it counts toward Flexible Savings Plans (FSP) across model providers[2]. Google also says customers can swap models for performance, latency and cost "without architectural lock-in or separate vendor contracts"[2].

Neither post publishes a price.

Google says customer prompts, context and outputs "remain entirely within your designated data perimeter and are never used to train foundation models," with Zero Data Retention options "available coming soon"[2]. It says the OpenAI models run on dedicated Google Cloud infrastructure, with Confidential Compute coming soon, and that queries are routed through isolated gateways so that data in transit stays inside Google Cloud and third-party model providers have no unauthorized access to the customer's cloud environment[2].

It also says every agent and model interaction runs under an authenticated identity, with VPC Service Controls and unified audit logging[2]. The posts do not attach a region list, a retention period or an outside audit to those sentences.

Yashodha Bhavnani, VP of AI Products at Box, is quoted in Google's post. She says that by using OpenAI models through Google Cloud, Box "will be able to give customers access to OpenAI's frontier AI models alongside Google Cloud's security and compliance capabilities"[2]. The quote is about what Box expects to offer. It does not say a Box product has switched over.

Greg Brockman, OpenAI's president and co-founder, said: "Google Cloud is a leading platform for developers and businesses building AI agents, writing software, and putting AI to work. This partnership is an important step in making our most capable models and Codex available where customers are already building, giving them even more choice in how they use frontier AI as we grow the agentic ecosystem together"[1].

Thomas Kurian, chief executive of Google Cloud, said: "We are focused on giving customers a broad selection of models, including those from OpenAI, so they can choose the right model for any given job, helping to deliver optimal quality and lower costs"[1].

OpenAI closes its post with one further sentence: "Microsoft remains OpenAI's primary cloud partner"[1].

The 21 October posts put OpenAI's models on a Google Cloud bill, inside a private preview, with Codex called out on both sides. Google's account of what a selected customer can do today is an API call and a local Codex CLI or Desktop connection into the Gemini platform. Its account of Antigravity, and OpenAI's account of general availability, are both still ahead. The figure the posts do give is Google's "more than 200" other models on the same platform. The figure they do not give is the price.