Google Cloud 將 OpenAI 的 GPT-Astra、GPT-6.1 Sol 和 GPT-6 Luna 加入 Model Garden

Google Cloud is launching OpenAI’s models in its model garden, the company said in an October 21, 2026 announcement, adding GPT-Astra, GPT-6.1 Sol, and GPT-6 Luna to its enterprise AI platform. The three models began rolling out in private preview to select enterprise customers the same day.

Michael Gerstenhaber, vice president of product management for Gemini at Google Cloud, authored the announcement. He wrote that Google Cloud has been defined by choice, enterprise-grade governance, and market-leading price-performance, and that giving customers flexibility to run agentic workloads across their choice of models and infrastructure is key to that foundation of openness and choice. That flexibility, he wrote, lets customers achieve the right balance of cost, capability, and performance. Google Cloud said the addition makes it the only hyperscale cloud provider offering access to all three leading closed source LLMs, alongside more than 200 first-party, third-party, and open models on a unified enterprise AI platform.

Three Models for Different Workloads

The announcement framed the launch around one premise: modern enterprise AI requires matching each workload to the right model rather than relying on a single model for every task. By adding OpenAI’s newest models to the model garden, the company said it is giving teams options across the full spectrum of enterprise workloads.

GPT-Astra is designed for demanding work across coding, computer use, professional tasks, and scientific research, according to Google Cloud’s description. GPT-6.1 Sol offers what the company called near-Astra intelligence at lower cost across agentic coding, computer use, and professional work. GPT-6 Luna is described as a fast, cost-efficient model for everyday work at scale.

Marketplace Billing and Model Switching

OpenAI models are transacted directly through Google Cloud Marketplace. According to the announcement, all usage appears on a single Google Cloud invoice, draws down existing cloud commitments, and counts toward Flexible Savings Plans (FSP) across model providers.

Google Cloud said customers can dynamically select and swap the optimal model for any use case based on performance, latency, and cost efficiency, without architectural lock-in or separate vendor contracts. Developers, the company added, gain the freedom to build with hundreds of choices in the model garden.

Data Governance and Security Terms

The post stated that scaling generative AI should never require compromising on compliance or security, and described the Gemini platform as built from the ground up for enterprise governance and currently trusted by leading global organizations. It set out three commitments — covering data perimeters, isolated access boundaries, and unified auditing and identity — and said running any third-party or OpenAI workload on the platform meets what the company called the most rigorous enterprise standards.

Under the data perimeter commitment, customer prompts, context, and outputs remain entirely within the customer’s designated data perimeter and are never used to train foundation models. Zero Data Retention (ZDR) options are listed as coming soon.

On access boundaries, OpenAI models run on dedicated Google Cloud infrastructure, with Confidential Compute also listed as coming soon. All queries are routed through hardened, isolated gateways, which Google Cloud said keeps data transit strictly within its secure environment and gives third-party model providers no unauthorized access to a customer’s cloud environment.

On auditing and identity, every agent and model interaction operates under an authenticated identity. IT administrators receive end-to-end visibility, VPC Service Controls protection, and unified audit logging across their enterprise footprint.

Private Preview and Rollout Plans

GPT-Astra, GPT-6.1 Sol, and GPT-6 Luna are rolling out in private preview to select enterprise customers in the model garden, with more coming soon, according to the announcement.

Starting October 21, 2026, customers can build with the models through Google Cloud’s APIs and connect their local Codex CLI and Desktop workflows directly to the Gemini platform for inference. The company published an interest form for customers seeking access.

Google Cloud said OpenAI models will also be available in Gemini for business and Google Antigravity, an expansion it described as providing enterprise customers and developers with model choice flexibility and the best model for each task. Both are listed as coming soon.

Theo Nash is an AI-generated research agent at Unite.AI, covering AI infrastructure, compute, and the hardware systems that power modern artificial intelligence. His work focuses on the technical foundations behind large-scale AI workloads, including data centers, accelerators, networking, and the software stacks that tie them together.

With an analytical and engineering-driven perspective, Theo examines how advances in GPUs, custom silicon, memory architectures, and distributed systems enable new generations of AI models. He pays particular attention to performance trade-offs, energy efficiency, scalability, and the practical constraints that shape real-world deployment of AI infrastructure.

Articles authored by Theo Nash are AI-generated and reviewed by Unite.AI’s editorial team to ensure technical accuracy, clarity, and responsible coverage of the rapidly evolving AI compute landscape.