Google Cloud is adding three OpenAI models to its platform, joining Gemini and Claude. The launch gives select enterprise customers another option for choosing AI models within Google’s cloud environment, with charges handled through a single bill.
Three OpenAI models, three stated roles
The new models are GPT-6 Astra, GPT-6.1 Sol, and GPT-6 Luna. Google describes Astra as suited to demanding work such as coding and research, Sol as a cheaper alternative with similar performance, and Luna as a fast, cost-efficient choice for everyday use.
Those descriptions frame the options around different needs: more demanding tasks, a lower-cost alternative, and routine use where speed and cost efficiency matter. The source does not provide performance figures or pricing, so customers will need to assess those details through the preview rather than infer them from the model labels.
Access starts with a private preview
The models are launching in private preview for select enterprise customers, and sign-ups are open now. They will also appear in Gemini for Business and Google Antigravity, extending access into those Google offerings.
For organizations already working with Google Cloud, the billing setup may make it simpler to account for use across the available models. Charges for OpenAI models will run through a single Google Cloud bill and count toward existing contracts. That consolidates billing, though it does not by itself establish how customers should choose among the models.
A broader model lineup on one platform
With Gemini, Claude, and OpenAI models available on one platform, Google Cloud now offers a mix of model providers to enterprise customers. Buyers can consider these options within the same cloud relationship, rather than treating access to each provider as an entirely separate platform decision.
Google says customer data won't be used for training. That assurance is relevant for enterprises evaluating where to send work, particularly when comparing models from different providers. The source does not describe other data-handling terms, so this statement should be understood specifically as a claim about training use.
The launch also puts Google in a distinctive position among major cloud providers: the article identifies it as the only major provider offering all three of Gemini, Claude, and OpenAI models. That breadth may give customers more choice, while making model selection and internal spending decisions more consequential.
The strategic tension for Google
Offering rival models can help Google Cloud serve customers who want access to OpenAI alongside Gemini and Claude. At the same time, the move exposes a tension within Google: compute capacity used to serve cloud customers also supports major competitors to Google’s own Gemini team.
The source reports that this resource conflict even pushed two-billion-dollar researcher Noam Shazeer back to OpenAI. The detail underscores how decisions about capacity can carry implications beyond customer access, reaching the competition for resources and people across AI efforts.
For enterprise buyers, the immediate change is a new set of OpenAI choices within Google Cloud, available through a private preview and consolidated billing. For Google, the expanded lineup balances customer demand for model choice against the resources its own AI team needs.