Provider deployment desk

Deploy open models
on Google Cloud.

Teams standardizing model serving in Vertex AI with custom containers and raw prediction.

Open Google Cloud ↗ Official docs ↗
25 model runbooks

Start with deployability, then compare quality

The rank is observed demand. The status is the hardware and orchestration burden on Google Cloud.

ModelPlanning floorRuntimeGoogle Cloud path
#1MiMo-V2.5Xiaomi · 310.8B 4×80GB400GB storage SGLanglatest Multi-GPU Vertex AI GPU deployment with at least 4x H100/A100 80GB-class accelerators
#2DeepSeek-V4-FlashDeepSeek · 284B 2×80GB250GB storage SGLanglatest Multi-GPU Vertex AI GPU deployment with at least 2x H100/A100 80GB-class accelerators
#3Hy3Tencent · 295B 4×80GB450GB storage SGLanglatest Multi-GPU Vertex AI GPU deployment with at least 4x H100/A100 80GB-class accelerators
#4GLM-5.2GLM · 753B 8×96GB1050GB storage vLLMlatest Multi-GPU Vertex AI GPU deployment with at least 8x H100/A100 80GB-class accelerators
#5DeepSeek-V4-ProDeepSeek · 1.6T 8×141GB1100GB storage SGLanglatest Multi-GPU Vertex AI GPU deployment with at least 8x H200-class accelerators
#6Nemotron 3 Ultra 550B-A55BNVIDIA · 550B 4×192GB450GB storage vLLMlatest Multi-GPU Vertex AI GPU deployment with at least 4x B200-class accelerators
#7MiniMax M3MiniMax · 427B 8×80GB600GB storage SGLanglatest Multi-GPU Vertex AI GPU deployment with at least 8x H100/A100 80GB-class accelerators
#8Step 3.7 FlashStepFun · 201B 1×120GB200GB storage vLLMlatest Single-node Vertex AI GPU deployment with at least 1x H200-class accelerators
#9Kimi K3Kimi · 2.8T 16×192GB3800GB storage SGLanglatest Cluster build At least 3072GB across multiple Google Cloud GPU nodes
#10MiMo-V2.5-ProXiaomi · 1.02T 8×141GB1400GB storage SGLanglatest Multi-GPU Vertex AI GPU deployment with at least 8x H200-class accelerators
#11gpt-oss-120bgpt-oss · 117B 1×80GB100GB storage vLLMlatest Single-node Vertex AI GPU deployment with at least 1x H100/A100 80GB-class accelerators
#12DeepSeek-V3.2DeepSeek · 685B 8×96GB950GB storage vLLMlatest Multi-GPU Vertex AI GPU deployment with at least 8x H100/A100 80GB-class accelerators
#13Gemma 4 31BGemma · 30.7B 1×80GB100GB storage vLLMlatest Single-node Vertex AI GPU deployment with at least 1x H100/A100 80GB-class accelerators
#14Nemotron 3 Super 120B-A12BNVIDIA · 120B 2×80GB200GB storage vLLMv0.18.1 Multi-GPU Vertex AI GPU deployment with at least 2x H100/A100 80GB-class accelerators
#15Gemma 4 26B-A4BGemma · 25.2B 1×80GB100GB storage vLLMlatest Single-node Vertex AI GPU deployment with at least 1x H100/A100 80GB-class accelerators
#16Kimi K2.6Kimi · 1.059T 8×80GB850GB storage vLLMlatest Multi-GPU Vertex AI GPU deployment with at least 8x H100/A100 80GB-class accelerators
#17Mistral NemoMistral · 12.2B 1×32GB100GB storage vLLMlatest Single-node Vertex AI GPU deployment with at least 1x 32GB+ accelerators
#18GLM-5GLM · 753.9B 8×96GB1000GB storage vLLMv0.19.0 Multi-GPU Vertex AI GPU deployment with at least 8x H100/A100 80GB-class accelerators
#19North Mini Code 1.0Cohere · 30B 1×80GB100GB storage vLLMlatest Single-node Vertex AI GPU deployment with at least 1x H100/A100 80GB-class accelerators
#20Laguna M.1Poolside · 225B 8×80GB650GB storage vLLMv0.21.0 Multi-GPU Vertex AI GPU deployment with at least 8x H100/A100 80GB-class accelerators
#21Laguna S 2.1Poolside · 118B 4×80GB350GB storage vLLMlatest Multi-GPU Vertex AI GPU deployment with at least 4x H100/A100 80GB-class accelerators
#22Kimi K2.5Kimi · 1.059T 8×80GB850GB storage vLLMlatest Multi-GPU Vertex AI GPU deployment with at least 8x H100/A100 80GB-class accelerators
#23Laguna XS 2.1Poolside · 33B 1×80GB100GB storage vLLMv0.21.0 Single-node Vertex AI GPU deployment with at least 1x H100/A100 80GB-class accelerators
#24MiniMax M2.7MiniMax · 229B 4×80GB350GB storage vLLMlatest Multi-GPU Vertex AI GPU deployment with at least 4x H100/A100 80GB-class accelerators
#25Kimi K2.7 CodeKimi · 1.059T 8×80GB850GB storage vLLMlatest Multi-GPU Vertex AI GPU deployment with at least 8x H100/A100 80GB-class accelerators
Agent-assisted setup

Hand this runbook to Claude Code or Codex

Open a terminal in the repository where you want the deployment files, start claude or codex, then paste this prompt. It asks the agent to verify sources and stop before it creates billable infrastructure.

Claude Code quickstart ↗ Codex docs ↗
Deployment prompt 
Prepare a reproducible open-model deployment project for Google Cloud; ask me to choose one of the models on this page before selecting GPU capacity. Use this guide as the starting context: https://www.getflops.ai/models/provider/gcp. Read the linked model card and provider documentation before choosing hardware or runtime settings. Open every linked primary source and flag any mismatch instead of guessing. Create a deployment folder containing README.md, .env.example with no secrets, a pinned start script or infrastructure manifest, and smoke-test.sh. Make the endpoint OpenAI-compatible where the runtime supports it. Run local/static validation, estimate the billable resources, and stop before provisioning paid infrastructure until I approve.
Guardrails included No secrets in files · verify primary docs · approval before spend
Provider source

Follow the current Google Cloud workflow

Google custom vLLM deployment docsA custom vLLM container in Vertex AI, backed by Artifact Registry ↗ Full launch boardCompare the same 25 models across all seven providers →