OpenAI's GPT 6 Models Land on Google Cloud and Antigravity

Google Cloud will host OpenAI's frontier models for the first time, a deal that puts GPT 6 and 6.1 into a Model Garden already stocked with more than 200 options and makes Google, by its own account, the only large cloud provider offering all three leading closed LLMs.

The arrangement, announced jointly by the two companies, carries a second, arguably weightier consequence: GPT models will appear in Antigravity, Google's agent-powered coding tool. For developers who have built workflows around Google's tooling but wanted OpenAI's models inside it, that door is now opening.

What exactly is shipping?

The full suite of current OpenAI models — GPT 6 and 6.1, spanning the variants Astra, Sol, and Luna — will launch on Google Cloud in the coming weeks, joining a Model Garden catalog that already includes Google's own Gemini family, open models such as Mistral, Llama, and Gemma, and Anthropic's Claude, which has been available there for some time.

Pricing will vary across the lineup. Astra, the largest of the three, will naturally cost the most to run, according to details accompanying the announcement. Sol and Luna sit below it, giving customers a spread of performance and price points from a single vendor console.

Google is also leaning on infrastructure it has already built. The company says its system supports dynamically moving workloads between LLMs based on performance, latency, and efficiency — meaning a customer isn't locked into one model per application but can shift traffic as conditions change.

What does this mean for Microsoft?

The obvious question is whether this dilutes Microsoft's long-standing position as OpenAI's cloud home. OpenAI moved to preempt it directly: the company notes that Microsoft continues to be its primary cloud partner. This is an expansion of distribution, in other words, not a defection.

But the competitive picture still shifts. Until now, enterprises wanting GPT models at scale went through OpenAI directly or through Azure. Adding Google Cloud as a sanctioned route gives buyers leverage they didn't have — three leading closed-model families, Gemini, Claude, and GPT, under one cloud billing relationship, plus a bench of open alternatives.

Why Antigravity matters as much as the cloud deal

The Model Garden addition is a distribution story. The Antigravity integration is a workflow story. Agent-powered coding tools increasingly function as the front door through which developers touch frontier models, and Antigravity until now has been tied to Google's model stack. Bringing GPT models into that surface means OpenAI's technology reaches developers who never open a Model Garden page at all.

That dynamic — models chasing developers inside the tools they already use — is where the platform competition is sharpest right now, and it's the part of this deal most likely to shape purchasing decisions over the next several quarters.

What comes next

The rollout begins in the coming weeks, with the Model Garden listing appearing first for enterprise customers and AI enthusiasts. Watch two things after launch: whether Google's dynamic workload-routing capability pulls real traffic between Gemini and GPT deployments, and whether Azure responds with pricing or bundling moves of its own. With all three closed-model leaders now on one cloud, the next battleground is less about access and more about cost per token and workflow lock-in.