Google Just Handed Developers Access to OpenAI, Claude, and Gemini in One Place

Google has done what once seemed unlikely. According to Ars Technica and The Decoder, the tech giant is bringing OpenAI's latest frontier models into Google Cloud and its Antigravity coding tool.

What changes for enterprise workflows?

For engineering teams, the addition means you no longer need separate billing or isolated vendor contracts to test competing models. Google's Model Garden now houses over 200 options, making it the only major cloud platform to officially host Gemini, Claude, and OpenAI's GPT lineup under one roof.

The integration runs entirely on Google's infrastructure. According to The Decoder, customer data will not be used for training, addressing a core privacy concern for corporate adopters.

How to access OpenAI models on Google Cloud

Enterprise customers can test the new setup immediately through a private preview rollout. Here is how teams can start using the feature:

• Submit an early access request directly through the Google Cloud console.
• Navigate to the Model Garden once approved to select between GPT-6 Astra, Sol, and Luna.
• Route API calls through your existing Google Cloud credits and billing dashboard.
• Integrate the models into supported clients like the Google Antigravity coding tool.

Is it worth changing your cloud stack?

If your company already maintains active spending commitments with Google Cloud, this move removes friction. You can substitute Gemini or Claude workloads with OpenAI's models without setting up new vendor pipelines. However, running heavy models like Astra will still burn through tokens quickly, requiring careful latency and cost monitoring.