Skip to content
Wednesday, October 21, 2026AboutContactRSS
Tech News

OpenAI launches GPT-6 models and Codex agent on Google Cloud platform

OpenAI and Google Cloud announced Wednesday that GPT-6 variants and the Codex coding agent are now available in private preview on the Google platform.

Illustration: A sleek, matte-black server chassis sits on a clean white desk.
AI-generated illustration: Threat Times

Key points

  • OpenAI’s GPT-6 Astra, GPT-6.1 Sol, and GPT-6 Luna are in private preview on Google Cloud’s Model Garden.
  • The Codex coding agent is also available, with general availability planned for the coming weeks.
  • This move places OpenAI models on all three major US cloud providers: AWS, Microsoft, and Google.

OpenAI and Google Cloud confirmed on Wednesday that OpenAI’s frontier models and its Codex coding agent are now available on the Google Cloud platform. The three new models, GPT-6 Astra, GPT-6.1 Sol, and GPT-6 Luna, are currently rolling out in a private preview within Google Cloud’s Model Garden, according to a statement from Google.

Select enterprise customers have access to these models first, with OpenAI planning to make them generally available in the coming weeks, the company said in a blog post. This expansion marks a significant shift in cloud hosting for OpenAI’s technology, placing its models on all three of the largest US cloud providers. Amazon Web Services began hosting OpenAI models in April after Microsoft’s exclusive license ended, and OpenAI added its agents to Amazon Bedrock last month, according to The Next Web.

Strategic Context and Market Position

Google now positions itself as the only hyperscale cloud provider offering all three leading closed-model AI systems, including Anthropic’s Claude and its own Gemini models, alongside OpenAI’s offerings. Greg Brockman, OpenAI’s president and co-founder, stated that the partnership makes their most capable models and Codex available where customers are already building applications. Thomas Kurian, chief executive of Google Cloud, added that the company is focused on providing a broad selection of models so customers can choose the right tool for any specific job.

Despite this new partnership, OpenAI stated that Microsoft remains its primary cloud partner. The deal comes as Google and OpenAI compete directly in both AI models and search services. Google recently released Gemini 4 Argon to paying users this month, intensifying the competitive landscape. Mike Leone, an analyst at Moor Insights & Strategy, told Axios that the deal demonstrates Google’s confidence, noting that Gemini now must win customers while OpenAI sits directly beside it in the marketplace.

Financial Implications and Infrastructure

Customers will purchase access to these models through the Google Cloud Marketplace, allowing them to use existing Google Cloud spending commitments to pay for usage, Google said. All usage appears on a single Google Cloud invoice, simplifying billing for enterprises. Regarding data privacy, Google stated that customer prompts and outputs remain inside the customer’s data perimeter and are never used to train models. The OpenAI models run on dedicated Google Cloud infrastructure, ensuring isolation from other workloads.

Financial data suggests high stakes for cloud providers hosting these models. Krystal Hu reported for Reuters that cloud providers typically take a cut of model sales on their platforms. According to Anthropic’s IPO filing, cited by Reuters, Anthropic generated 47% of its 2025 revenue through Amazon’s and Google’s cloud platforms. Google’s cloud revenue growth accelerated to about 82% in the second quarter, as reported by CNBC’s Jordan Novet, compared to 37% for Amazon and 42% for Microsoft.

What should you do?

Security and technology professionals should monitor their cloud environments for new API endpoints associated with OpenAI models as they move to general availability. Organizations using Google Cloud should review their data perimeter configurations to ensure that prompts and outputs remain isolated according to Google’s stated privacy guarantees. Teams should also evaluate whether the new GPT-6 variants offer specific advantages over existing Gemini or Claude models for their particular workloads.

What to do and how to stay safe: OpenAI

  • Review current cloud spending commitments to determine if they can be applied to new OpenAI model usage on Google Cloud.
  • Verify that data isolation protocols are correctly configured for any new AI integrations to ensure prompts stay within your data perimeter.
  • Test the private preview versions of GPT-6 Astra, Sol, and Luna in non-production environments to assess performance before general availability.
  • Monitor official announcements from OpenAI and Google Cloud for the exact timeline of general availability and any changes to pricing structures.

Step-by-step guide: Mobile Device Security: How to Protect Data Beyond Passwords

General security guidance from the Threat Times newsroom. It is not confirmed advice from the organisations named in this story.

Frequently asked questions

Which OpenAI models are available on Google Cloud?

GPT-6 Astra, GPT-6.1 Sol, and GPT-6 Luna are currently in private preview, along with the Codex coding agent.

When will these models be generally available?

OpenAI plans to make the models generally available in the coming weeks, following the current private preview for select enterprise customers.

Does Google use customer data to train these models?

No, Google stated that customer prompts and outputs stay inside the customer’s data perimeter and are never used to train models.

Sources

  1. The Next Web
Related articles