Google Cloud and OpenAI announced on 21 October that GPT-6 Astra, GPT-6.1 Sol and GPT-6 Luna are being added to Google's Model Garden. Access starts as a private preview for selected enterprise customers through the API, and Codex CLI and Desktop workflows can also run on Google infrastructure. Billing goes through Google Cloud Marketplace and counts toward customers' existing spending commitments. Google says data stays within the customer's perimeter and is not used for training, while zero data retention options will follow later.

What it means: The deal shows how much the enterprise AI map has shifted. Until recently OpenAI was tied mainly to Microsoft, but since that exclusivity ended its models are distributed by AWS and now by Google as well. For customers this decouples the choice of model from the choice of cloud: a company that already keeps its data and contracts with Google can try GPT, Gemini and Claude from one place, with a single invoice and the same access controls. For Greek businesses and software houses building agent-based applications, the practical gain is being able to compare cost and quality without a new procurement process. That flexibility also demands discipline: teams need a record of which model is used for which task, where data is processed and which terms apply, especially when personal data or AI Act obligations are involved. It is also worth noting that some safeguards, such as confidential compute, are not available yet, so sensitive workloads deserve caution until they are switched on.

Read the original on Google Cloud Blog

Related service: Custom software

Related stories