Ars Technica: OpenAI GPT 6.1 Models Head to Google Cloud and Antigravity

OpenAI and Google have reached an agreement that will bring the former’s frontier models to Google Cloud services and the Antigravity coding tool.

Three GPT 6.1 models enter Model Garden. The full suite of current GPT 6.1 models—Astra, Sol, and Luna—will launch on Google Cloud in the coming weeks. They join more than 200 options already available in the Model Garden, which includes Google’s Gemini models plus Mistral, Llama, Gemma, and Anthropic’s Claude.

Google states it is now the only large cloud provider offering access to all three leading closed LLMs. OpenAI notes that Microsoft remains its primary cloud partner.

Workloads run on Google infrastructure. The OpenAI models will run entirely on Google’s systems, with transactions handled through the Google Cloud Marketplace. Enterprise customers with existing Google Cloud spending commitments can switch to the GPT models without new billing setups.

Astra will carry the highest cost to run, yet the platform supports moving workloads between models based on performance, latency, and efficiency.

Private preview begins for enterprises. All three OpenAI LLMs are now rolling out in private preview for select enterprise customers. Interested users can submit a request for early access through the API, including via Codex CLI or Codex Desktop.

Antigravity gains the same models. The GPT models will also appear in Google’s Antigravity agent-powered coding tool. Until now Antigravity has supported Gemini and Claude; Google has not yet specified whether the full trio of OpenAI models will be included.