OpenAI's GPT-Astra and GPT-6 models, including GPT-6 Astra, are now available on Google Cloud's Model Garden, which the report says completes OpenAI's availability across the big three cloud providers.

OpenAI will begin selling its frontier models and its Codex coding agent through Google Cloud within weeks, the two companies said Wednesday. The deal puts the ChatGPT developer's technology on all three of the largest U.S. cloud platforms and further erodes the wall between two of AI's most direct competitors.

Under the arrangement, OpenAI's latest models, including GPT-6 Astra, will appear in Google Cloud's Model Garden and be purchasable through the Google Cloud Marketplace. Customers will be able to draw on existing cloud spending commitments to pay for them, according to a company blog post. The models will also reach Google's Antigravity development environment and Gemini Agent, with general availability planned in the coming weeks.

The tie-up gives companies in regulated industries a route to OpenAI's models through Google Cloud's access controls, data governance tools and infrastructure, which the companies said would help customers build so-called agentic AI workflows.

For OpenAI, the agreement is the latest step in a distribution campaign that began after its exclusive licensing arrangement with Microsoft lapsed in April. Amazon Web Services moved the following day to host OpenAI models, and Amazon has since widened the lineup available through its Bedrock platform. Neither company has disclosed how much additional business that arrangement has generated.

The Google deal adds a layer of complexity to a relationship that has been adversarial for most of its history. OpenAI was founded in 2015 partly as a counterweight to the concentration of AI research at Google, and the two now compete in core models and in search. Yet the capital and computing demands of building and running large models have pushed them toward cooperation. Last year, OpenAI said it was using Google's cloud to run ChatGPT and its application programming interface, and Google struck a cloud agreement to supply OpenAI with additional computing capacity.

Google has been rebuilding its cloud business around that openness. Annualized revenue growth in the segment accelerated to roughly 82% in the second quarter from 63% in the first, helped by demand for AI models and custom chips, according to CNBC. Amazon and Microsoft posted second-quarter cloud growth of 37% and 42%, respectively. All three report quarterly results next week.

Thomas Kurian, chief executive of Google Cloud, has described the platform as an open one that welcomes any model popular with developers. That position lets Google sell a rival's products alongside its own Gemini line. Google unveiled its next-generation Gemini 4 Argon model in September.

"For Google, I read the deal as confidence, because Gemini now has to win those customers with OpenAI sitting right next to it," Mike Leone, vice president and analyst at Moor Insights & Strategy, told Axios. He added that Google is unlikely to see Gemini cannibalized, since customers pay for its cloud platform, data services and security regardless of which models they run.

OpenAI framed the move as a validation of Google's infrastructure. "Google Cloud's longstanding investment in machine learning and AI has made it a destination for AI workloads and the developers and data scientists building them," the company said in its blog post.

The arrangement reflects a broader shift in which hyperscalers have become the primary commercial channel for frontier AI labs, typically taking a cut of sales made through their platforms. Anthropic, the OpenAI rival backed by Amazon and Google, drew 47% of its 2025 revenue from those two cloud platforms, according to its IPO filing as reviewed by Reuters.

Distribution, however, is only part of the contest. OpenAI must still persuade enterprises to pick its models over those from Google, Anthropic and others that sit on the same shelves. That raises the value of the cloud providers' existing corporate relationships and infrastructure.

"Once every cloud has the top models, I think the competitive advantage goes to whoever does the best job with a company's data, governance, and costs," Leone said.

The stakes are rising on both sides. Most business spending on AI flows to OpenAI and Anthropic, according to data from corporate card issuer Ramp, though some corporate IT departments have folded Gemini into their systems. OpenAI, meanwhile, is courting investors at a $1.4 trillion valuation and targeting an initial public offering.

For Google, the deal captures enterprise workloads that might otherwise have gone to AWS or Microsoft Azure, and it deepens a relationship that began with a computing contract. For OpenAI, it removes a reason for corporate buyers to maintain a separate vendor relationship just to reach its models.

**Deal at a glance**

- **Models covered:** GPT-6 Astra and other frontier models, plus the Codex coding agent

- **Sales channels:** Google Cloud Model Garden and Google Cloud Marketplace

- **Payment mechanism:** Existing Google Cloud spending commitments

- **Additional surfaces:** Antigravity development environment, Gemini Agent

- **Availability:** General availability planned in the coming weeks

- **Cloud coverage:** Google Cloud, AWS, Microsoft Azure

*Details as announced by the companies on Wednesday.*