Skip to content
Trending
AI2 min read

Google Cloud Lets Enterprise Customers Run GPT-6 Alongside Gemini

Three new OpenAI models join a private preview on Google Cloud, letting businesses pick GPT-6 Astra, Sol or Luna without opening a second contract with OpenAI.

By Model Card

The AI Desk · (1 hour ago)

Reported fromoutlets, credited2
Google Cloud logo displayed alongside icons representing Gemini, Claude and OpenAI model offerings
Google Cloud logo displayed alongside icons representing Gemini, Claude and OpenAI model offeringsPhoto via The Decoder

Google Cloud started offering OpenAI's models to enterprise customers on Tuesday, a move that makes it the only major cloud platform where businesses can run Gemini, Claude and OpenAI's models from a single account.

Three models are available in the private preview that opened today. GPT-6 Astra is pitched at demanding work like coding and research. GPT-6.1 Sol is described as a cheaper option with similar performance. GPT-6 Luna is the fast, low-cost pick for everyday tasks. Sign-ups are open now, though access is limited to select enterprise customers rather than the general public.

The models will also show up inside Gemini for Business and Google's Antigravity platform, so customers already working in those tools won't need to switch environments to call on OpenAI's models. Every charge, whatever the model, lands on one Google Cloud bill and counts toward contracts customers already have in place. Google says it won't use customer data from these sessions to train its own models, a reassurance aimed at enterprise clients wary of handing proprietary data to a rival AI lab through the back door.

It's an odd arrangement on its face. Google Cloud is the infrastructure arm of a company that builds Gemini, its own answer to OpenAI's models. Selling compute to run a direct competitor's product looks like helping the other side.

But cloud providers have increasingly separated the infrastructure business from the model business. Microsoft Azure has hosted OpenAI's models for years while Microsoft also ships its own Copilot products. Amazon's AWS offers Claude through Bedrock alongside its in-house Titan models. Google had held back from doing the same with OpenAI specifically, until now.

The Decoder reports this creates a recurring tension inside Google: capacity spent serving outside customers running OpenAI's or Anthropic's models is capacity not spent scaling Gemini. That conflict reportedly played a part in pulling researcher Noam Shazeer, whom the outlet describes as a two-billion-dollar hire, back to OpenAI after a stint at Google.

For enterprise customers, the pitch is simpler: pick whichever model fits a given job, Gemini for one task, Claude for another, GPT-6 Astra for a third, without juggling separate contracts, separate billing and separate data agreements with three different companies.

Google hasn't said when the OpenAI models might move beyond private preview to general availability on Google Cloud, or whether pricing will match what OpenAI charges directly through its own API. Those details, along with how long the preview list stays limited to select customers, are the next things to watch.

Reported from

How this story was made

Written by the Hitechreports desk from the reporting credited above, with facts attributed to their original publishers. We do not test devices ourselves; anything about performance, battery life or cameras comes from the outlets that did. Prices are as reported at the time of writing. Editorial policy · Report an error

What did you make of it?

Filed by The AI Desk

Models, assistants and the companies and chips behind them, reported from what was released and what was claimed, with the difference kept clear.

More from this desk →

Be the first to comment

Join the argument. No password, just your email or a passkey.

    Read next

    More AI →