GMI Cloud

What's good about GMI Cloud

• Transparent published per-GPU hourly rates
• Access to current Blackwell-generation systems
• OpenAI-compatible inference API alongside dedicated GPU clusters

GMI Cloud GPU prices

4 GPU models across 4 configurations. Rank compares GMI Cloud's cheapest on-demand listing for each GPU with every provider's cheapest.

What a run costs on GMI Cloud

24 hours at today's cheapest listing for each GPU, at the size it is listed.

• 1× H100 SXM for 24 hours

  At $2.60/GPU-hr

  $62.40

  $18.96 under the market median rate

• 1× H200 SXM for 24 hours

  At $3.20/GPU-hr

  $76.80

  $31.28 under the market median rate

• 1× B200 for 24 hours

  At $5.00/GPU-hr

  $120.00

  $70.56 under the market median rate

The market median is every provider's cheapest on-demand price for that GPU.

GMI Cloud LLM API prices

Pay-per-token pricing, per 1M tokens.

Prices last updated November 13, 2026

Billing & terms

On-Demand Containers

Hourly billing for self-serve GPU containers

Reserved Private Cloud

Discounted longer-term reservations of dedicated GPU clusters

Serverless Inference

Per-token billing for LLM endpoints and per-request billing for image and video models

Regions

Data centers in North America and Asia, with region-aware pricing and unified billing

Support

Documentation, self-service console, Discord community, and enterprise support via sales

Watch out for

• Smaller global footprint than hyperscalers
• Newer entrant relative to long-established providers
• Published GPU catalog covers only datacenter-class parts, with no lower-tier options such as L4 or L40S

Features

Serverless Inference

OpenAI-compatible endpoints for LLM and multimodal models with request batching and scaling to zero

Dedicated GPU Compute

Managed Kubernetes clusters, container instances, and bare-metal servers with RDMA-ready networking

Blackwell Capacity

GB200 available and GB300 on pre-order alongside H100, H200, and B200 systems

Studio and AgentBox

Visual workflow builder for multi-step model pipelines plus a marketplace for publishing and using AI agents

Reserved + On-Demand

Both hourly on-demand capacity and longer-term committed reservations are published

Getting started

1. Step 1

   Create an account

   Sign up for the GMI Cloud console

2. Step 2

   Pick a GPU and region

   Select an on-demand container, bare-metal cluster, or inference endpoint

3. Step 3

   Deploy your workload

   Launch via the console or programmatically through the GMI API