OpenAI and Google have reached an agreement to make OpenAI’s latest frontier models available on Google Cloud services and within Google’s agent-powered coding tool, Antigravity. The GPT 6.1 model suite—comprising Astra, Sol, and Luna—is rolling out in private preview for enterprise cloud customers and will be accessible via Google’s Model Garden and Marketplace.

Despite Microsoft remaining OpenAI’s primary cloud partner, the arrangement allows Google Cloud to run GPT inference entirely on its own infrastructure. Google stated that this integration makes it the only large cloud provider offering unified access to all three leading closed large language model suites: Gemini, Claude, and GPT.

In addition to cloud deployment, the GPT models will be integrated into the Google Antigravity developer client alongside existing options. Google’s cloud system will support dynamic workload routing between models based on latency, performance, and efficiency requirements.

Why it matters

  • Cloud lock-in weakens as multi-model deployments become standard, allowing enterprise buyers to route tasks dynamically across top-tier foundation models.

  • Google Cloud expands its competitive edge by offering the three leading proprietary LLM families (Gemini, Claude, GPT) on a single platform.

  • Developers gain broader model choice in coding environments like Antigravity, decreasing reliance on single-vendor developer tools.

Source: arstechnica.com