OpenAI GPT-6 Models Come to Google Cloud

OpenAI and Google fight over chatbots, search and developers’ attention. From today they also share a sales channel. Google Cloud has started offering OpenAI’s GPT-6 Astra, GPT-6.1 Sol and GPT-6 Luna in its model garden, the catalogue of models that business customers can call through Google’s own platform.

The catch for most companies is access. Google’s announcement, by Michael Gerstenhaber, says the models are “rolling out now in private preview to select enterprise customers”, with an interest form for everyone else. OpenAI’s own post says general availability is planned “in the coming weeks”. For a company that runs most of its systems on Google, though, the deal means OpenAI’s models can soon be bought without opening a separate account with OpenAI.

Google is selling the deal on billing and control

The pitch from Google is less about the models than the paperwork around them. Usage is “transacted directly through Google Cloud Marketplace”, so it appears on a single Google Cloud invoice, draws down existing cloud commitments and counts towards Google’s Flexible Savings Plans. For a customer that has already committed to spend a set amount with Google, OpenAI usage now counts towards that total.

Google describes the three models in a ladder. Astra is aimed at “demanding work across coding, computer use, professional tasks, and scientific research”. Sol offers “near-Astra intelligence at lower cost”, and Luna is “a fast, cost-efficient model for everyday work at scale”. Developers can reach them through Google’s APIs and, Google says, can point their local Codex CLI and Desktop workflows at its Gemini platform for inference.

On data, Google says customer prompts, context and outputs stay within the customer’s chosen perimeter and are “never used to train foundation models”. Two of the stronger protections are not live yet: Zero Data Retention options and dedicated infrastructure with Confidential Compute are both listed as “coming soon”. So is support for OpenAI models inside Gemini for business, which OpenAI’s post calls the Gemini Agent, and inside Google Antigravity, Google’s agentic coding tool.

Google also uses the launch to make a competitive claim. It says it is “now the only hyperscale cloud provider offering access to all three leading closed source LLMs”, alongside more than 200 other models. The post does not name the three, and the claim is Google’s.

Microsoft’s exclusivity ended in April, and the clouds moved quickly

The deal follows a change in OpenAI’s arrangement with Microsoft. Reuters reported that OpenAI’s exclusive distribution arrangement with Microsoft expired in April, and that OpenAI has since expanded to Amazon Web Services and now Google Cloud. OpenAI’s post ends by noting that “Microsoft remains OpenAI’s primary cloud partner.”

Reuters framed the deal as part of OpenAI’s push to sell to more enterprise customers. This week it also emerged that Emirates NBD will run OpenAI models on in-country inference in the UAE. Cloud providers usually take a cut of sales through their platforms, Reuters noted.

Both companies framed it as customer choice. “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,” Google Cloud chief executive Thomas Kurian said in OpenAI’s post. Reuters pointed to how much frontier labs already rely on the clouds: Anthropic, which Google and Amazon both back, made 47% of its 2025 revenue through those two companies’ cloud platforms, according to its IPO filing.

Neither company has said what the models will cost on Google Cloud, or which regions the preview covers. Those answers should come with general availability, which OpenAI has put “in the coming weeks”. The next test is whether the promised zero-retention and Confidential Compute options arrive at the same time, since regulated buyers are the customers both companies say they are chasing.