OpenAI’s GPT-6 Models Reach Google Cloud in Private Preview

Google data center buildings in Council Bluffs, Iowa
A Google data center in Council Bluffs, Iowa. Photo: Chad Davis via Wikimedia Commons, CC BY 2.0

OpenAI’s newest models are now on Google Cloud. GPT-6 Astra, GPT-6.1 Sol and GPT-6 Luna began rolling out in private preview to select enterprise customers on Oct. 21, and OpenAI says it plans general availability “in the coming weeks.”

The deal matters most to companies that already run on Google’s cloud. They can now buy OpenAI models through the same contract and invoice as the rest of their Google spending. It also means OpenAI’s models are sold on all three of the biggest US cloud platforms, a point CNBC noted in its report on the announcement.

Three GPT-6 models and Codex are in the first wave

Google’s announcement, written by Gemini product vice president Michael Gerstenhaber, lists three OpenAI models in its Model Garden catalog. GPT-6 Astra is the top model, built for coding, computer use, professional tasks and scientific research. GPT-6.1 Sol offers what Google calls “near-Astra intelligence at lower cost.” GPT-6 Luna is the fast, cheap option for high-volume everyday work.

The list skips GPT-6.1 Astra. OpenAI shelved that model in September after safety tests flagged deceptive behavior, so Google customers get the earlier Astra release.

Developers can also point OpenAI’s Codex coding agent at Google from day one. Google says users can connect their local Codex command-line and desktop workflows to its Gemini platform for inference. That covers the tool OpenAI rebuilt as a standalone desktop app in February.

OpenAI models will come later to Google’s Antigravity coding tool and to Gemini for business, which OpenAI’s post calls the Gemini Agent. Neither company gave a date.

Customers who want in now have to apply. Both companies are taking requests through interest forms, so the preview is not open to every Google Cloud account.

Customers can pay with money they already committed to Google

The commercial terms are the real selling point. Google sells the OpenAI models through Google Cloud Marketplace, so all usage shows up on a single Google Cloud invoice.

Large companies often sign multi-year spending commitments with a cloud provider in exchange for discounts. Google says OpenAI purchases will draw down those commitments and count toward its Flexible Savings Plans. In practice, a finance team with unspent Google credit can now spend it on GPT-6 instead of signing a separate contract with OpenAI.

That matters because marketplace sales already drive much of the AI model business. Anthropic disclosed in its IPO filing that cloud marketplaces generated 47% of its 2025 revenue, as Techstrong reported. OpenAI is courting investors at a $1.4 trillion valuation ahead of its own planned listing, according to CNBC. It now has access to the same channel.

Some of Google’s privacy promises are still “coming soon”

Google’s pitch leans on governance for banks, hospitals and other regulated buyers. It says prompts, context and outputs stay inside a customer’s designated data perimeter and are never used to train foundation models. OpenAI models run on dedicated Google Cloud infrastructure, and queries pass through isolated gateways. Google also offers its existing identity controls, VPC Service Controls and unified audit logs.

Two features that compliance teams often ask for are not ready yet. Google lists both Zero Data Retention options and Confidential Compute as “coming soon.” Buyers with strict retention rules should treat the current preview as a test environment until those ship.

Box is the one customer Google named. “By using OpenAI models through Google Cloud, we will be able to give customers access to OpenAI’s frontier AI models alongside Google Cloud’s security and compliance capabilities,” said Yashodha Bhavnani, Box’s vice president of AI products, in Google’s post.

Neither company disclosed pricing or revenue-sharing terms.

Microsoft’s lost exclusivity opened the door

Until this spring, Microsoft’s contract stood in the way. Microsoft held an exclusive license to OpenAI’s technology until late April 2026, when the two companies rewrote their agreement. Amazon moved within a day, and AWS unveiled OpenAI models and Codex on Bedrock in limited preview.

Google is the last of the big three to sign on, nearly six months later. OpenAI’s own post ends with a reassurance for its biggest backer: “Microsoft remains OpenAI’s primary cloud partner.”

The two companies were already linked. OpenAI said last year that it uses Google’s cloud to run ChatGPT and its API, CNBC reported. This deal reverses that flow, with Google now selling OpenAI’s models to its own customers.

“This partnership is an important step in making our most capable models and Codex available” where customers already build, OpenAI President Greg Brockman said, as quoted by The Next Web. Google Cloud CEO Thomas Kurian framed it as model choice: “We are focused on giving customers a broad selection of models, including those from OpenAI.”

Gemini now shares a shelf with its biggest rival

Google now claims to be the only major cloud that sells all three leading closed model families: its own Gemini, Anthropic’s Claude and OpenAI’s GPT. Model Garden lists more than 200 first-party, third-party and open models.

The bet is that Google wins either way. If a customer picks GPT-6 over Gemini, the compute and the contract still run through Google Cloud. The risk is that Gemini loses head-to-head contests on Google’s own platform. Moor Insights & Strategy analyst Mike Leone put it plainly to Axios: “Gemini now has to win those customers with OpenAI sitting right next to it.”

Google has momentum to defend. Its cloud revenue growth accelerated to about 82% in the second quarter, compared with 37% at Amazon and 42% at Microsoft, according to CNBC. All three companies report quarterly results next week, the first chance for executives to say how much demand the OpenAI preview is drawing.

Exit mobile version