Get started

GPUs.io API

Programmatic access to live cloud-GPU prices, historical trends, and the full GPUs.io catalogue of GPU types and providers — the same data behind our comparison tables, served as clean JSON.

Base URL

https://api.gpus.io/v1

Version

v1 · stable

Auth

Bearer API key

Format

JSON · UTF-8

Private beta. The API is currently in private beta and authenticated with an API key — there is no public pricing yet. See Rate limits & access for the current limits.

Quickstart

Every request is a plain HTTPS call with your key in the Authorization header. Here are the two cheapest live H100 SXM offerings, on-demand:

curl -G https://api.gpus.io/v1/prices \
  -H "Authorization: Bearer $GPUS_API_KEY" \
  --data-urlencode gpu=h100-sxm \
  --data-urlencode rentalType=on_demand \
  --data-urlencode limit=2

A successful call returns a data array and a pagination object:

200 OK

{
  "data": [
    {
      "gpu": {
        "key": "h100-sxm",
        "name": "NVIDIA H100 SXM",
        "vramGb": 80
      },
      "provider": {
        "id": "lambda",
        "name": "Lambda",
        "website": "https://lambdalabs.com"
      },
      "rentalType": "on_demand",
      "commitmentTermMonths": null,
      "gpuCount": 8,
      "pricePerGpuHourUsd": 1.99,
      "totalPricePerHourUsd": 15.92,
      "specs": {
        "vcpu": 208,
        "ramGb": 1800,
        "bootDiskGb": 1024,
        "scratchDiskGb": 22000
      },
      "regions": [
        "us"
      ],
      "available": true,
      "lastUpdated": "2026-06-03T08:12:04Z"
    },
    {
      "gpu": {
        "key": "h100-sxm",
        "name": "NVIDIA H100 SXM",
        "vramGb": 80
      },
      "provider": {
        "id": "runpod",
        "name": "RunPod",
        "website": "https://runpod.io"
      },
      "rentalType": "on_demand",
      "commitmentTermMonths": null,
      "gpuCount": 1,
      "pricePerGpuHourUsd": 2.49,
      "totalPricePerHourUsd": 2.49,
      "specs": {
        "vcpu": 24,
        "ramGb": 125,
        "bootDiskGb": 80,
        "scratchDiskGb": 1024
      },
      "regions": [
        "us",
        "nl"
      ],
      "available": true,
      "lastUpdated": "2026-06-03T08:12:04Z"
    }
  ],
  "pagination": {
    "nextCursor": "eyJpZCI6MTQ4Mn0",
    "limit": 2
  }
}

Endpoints

Four read-only endpoints cover the whole dataset. Prices are the live feed; the reference lists give you the valid filter values.

GETList prices Current price for every GPU · provider · configuration. The core comparison feed.GETPrice history Historical price series for a single GPU type, built on our trend data.GETList GPU types The catalogue of GPU types — valid gpu filter values plus hardware specs.GETList providers Tracked cloud providers, with descriptions and websites.

Conventions

All endpoints are read-only GET requests. There is no write surface.
Prices are per GPU, per hour, in USD unless a field name says otherwise (e.g. totalPricePerHourUsd is for the whole configuration).
Timestamps are ISO 8601 in UTC (e.g. 2026-06-03T08:12:04Z).
Regions are ISO 3166-1 alpha-2 country codes (e.g. us, de), not provider-specific zone names.
List responses are cursor-paginated — see Pagination & errors.
Identifiers like gpu.key and provider.id are stable strings you can safely store and join on.

Next steps

Authentication Get a key and authorize requests.Pagination & errors Page through results and handle failures.Rate limits Limits, headers, and access.

Main site Access Support