Google Cloud
Google Cloud is Google's cloud platform. For LLMs it offers ready-to-use open models through Vertex AI Model Garden, GPU and TPU virtual machines on Compute Engine, and managed Kubernetes on GKE for running models at scale.
Overview
Google Cloud is Google's public cloud, with data centres in more than 40 regions across the Americas, Europe, the Middle East, Africa and Asia Pacific.
It gives you three main ways to run large language models. Vertex AI Model Garden offers open models such as Llama, DeepSeek, Qwen, gpt-oss and Google's own Gemma, either as fully managed APIs billed per token or deployed in a few clicks to your own endpoint. Compute Engine provides NVIDIA GPU and Google TPU virtual machines for teams that want full control. Google Kubernetes Engine (GKE) runs and scales model servers such as vLLM on clusters of those GPUs and TPUs.
Google Cloud suits teams already using Google's data and AI tools, those who want access to TPUs, and businesses that need to keep data in a specific region.
Plans from Google Cloud
| Hosting plan | Hosting type | Pricing range | Model families |
|---|---|---|---|
| Model Garden | Serverless API | US$0.07 – US$5.4 | Llama, Mistral, DeepSeek, Qwen, gpt-oss, Gemma |
| Google Kubernetes Engine (GKE) | GPU cloud | From US$0.35 | gpt-oss, Qwen, DeepSeek, Llama, Gemma, Mistral |
| Compute Engine | GPU cloud | From US$0.35 | gpt-oss, Qwen, DeepSeek, Llama, Gemma, Mistral |