OpenAI will sell its frontier models and its Codex coding agent through Google Cloud in the coming weeks, the companies said on Wednesday, a deal that puts OpenAI's technology on all three of the biggest American cloud platforms and deepens ties between two of the fiercest rivals in artificial intelligence.
Google Cloud announced that OpenAI's newest models are launching in its Model Garden, the catalogue of more than 200 first-party, third-party and open models it offers to enterprise customers. Google said it is now the only hyperscale cloud provider offering access to all three leading closed-source model families — its own Gemini, Anthropic's Claude and OpenAI's GPT — on a single platform.
Three models lead the rollout. GPT-6 Astra is positioned for demanding work across coding, computer use, professional tasks and scientific research. GPT-6.1 Sol offers near-Astra intelligence at lower cost for agentic coding, computer use and professional work. GPT-6 Luna is the fast, cost-efficient option for everyday work at scale. The models are rolling out now in private preview to select enterprise customers, according to Google, with OpenAI saying general availability is planned in the coming weeks. Customers can build through Google's APIs and, starting today, connect their local Codex CLI and desktop workflows directly to Google's Gemini platform for inference. OpenAI models are also coming soon to Gemini for business and to Google Antigravity, Google's coding environment.
For enterprise buyers, the purchasing mechanics may matter as much as the models. Because the models are transacted directly through Google Cloud Marketplace, usage appears on a single Google Cloud invoice, draws down existing cloud spending commitments and counts toward Flexible Savings Plans across model providers. That removes a separate vendor contract for companies already committed to Google Cloud, and it lets teams swap models based on performance, latency and cost without what Google calls architectural lock-in.
Google is also pitching its governance controls as a reason to buy OpenAI models through its cloud rather than directly. It says customer prompts, context and outputs remain within a designated data perimeter and are never used to train foundation models, with Zero Data Retention options coming soon. OpenAI models run on dedicated Google Cloud infrastructure, with Confidential Compute coming soon, and every agent and model interaction operates under an authenticated identity with unified audit logging and VPC Service Controls protection. Reuters noted the arrangement will also let customers in regulated industries use those access controls and data governance tools to build agentic workflows.
The deal completes a rapid unwinding of the old distribution map. OpenAI's exclusive distribution arrangement with Microsoft expired in April, Reuters reported. OpenAI then expanded distribution to Amazon Web Services, and Google Cloud is the latest hyperscaler to join the list. The two new partners still compete directly in AI products and search, and OpenAI was founded in 2015 partly as a counterweight to the concentration of AI research at companies like Google — which makes Google selling GPT models by the token a notable turn.
It is not the first cooperation between them. Google previously struck a cloud deal to provide OpenAI with additional computing capacity last year, and Google Cloud chief executive Thomas Kurian has described the platform in a previous Reuters interview as an open platform that welcomes any model popular with developers. Wednesday's announcement follows that logic to its conclusion: in a market where enterprises want choice more than loyalty, the cloud that stocks every leading model wins the customer, even when one of those models belongs to a rival.