Offers - DeployGPU

Offers are GPU instances available for deployment. These endpoints are public and do not require authentication.

List Offers

GET /api/v1/offers

Search and filter available GPU offers. Results are sorted by price per GPU (ascending) by default.

Query Parameters

Param	Type	Default	Description
`gpu`	string	none	Comma-separated GPU slugs (for example, `h100-sxm-80gb,a100-pcie-80gb`). See GPU Families for available slugs
`provider`	string	none	Comma-separated provider slugs (for example, `verda,hyperstack`)
`continent`	string	none	Comma-separated (for example, `Europe,North America`)
`min_gpu_count`	int	none	Minimum GPU count per instance
`max_gpu_count`	int	none	Maximum GPU count per instance
`sort_by`	string	`pricePerGpu`	Sort field: `pricePerGpu`, `priceHourly`, `gpuCount`
`sort_order`	string	`asc`	`asc` or `desc`
`page`	int	0	Page number (0-indexed)
`size`	int	25	Page size (max 100)

Response

{
  "offers": [
    {
      "id": "c0858d33-a3de-4fbd-8fee-98d009c3ec7d",
      "provider": {
        "name": "VERDA",
        "slug": "verda",
        "maintenance": false
      },
      "gpu": {
        "name": "NVIDIA A100 SXM4 80GB",
        "slug": "a100-sxm-80gb",
        "vramGB": 80,
        "architecture": "Ampere"
      },
      "gpuCount": 1,
      "region": "FIN-HEL",
      "regionInfo": {
        "countryCode": "FI",
        "countryName": "Finland",
        "cityState": "Helsinki",
        "continent": "Europe"
      },
      "priceHourly": 1.29,
      "pricePerGpu": 1.29,
      "isAvailable": true,
      "specs": {
        "vcpu": 22,
        "ram": 120,
        "disk": null
      },
      "storageCostPerGB": 0.2,
      "storageMinGB": 1,
      "storageMaxGB": 500,
      "costDataComplete": true,
      "deploymentType": "virtual",
      "nvlinkEnabled": true,
      "interconnect": "SXM"
    }
  ],
  "page": 0,
  "totalPages": 44,
  "totalElements": 88,
  "hasMore": true
}

Understanding provider.maintenance

provider.maintenance is true while the offer’s provider is in maintenance. The offer stays in the list, but creating a deployment from it returns 503 provider_maintenance until the maintenance ends. Skip those offers, or look up the expected end on GET /api/v1/providers.

Understanding deploymentType

deploymentType is virtual for a VM offer on QuantaCloud, Massed Compute, Verda or Hyperstack, and dedicated for a Vast.ai container offer. Both kinds appear in this list together. Only virtual offers can be created through POST /api/v1/deployments. A dedicated offer_id there returns provider_not_supported. Filter on this field before you try to create anything.

Understanding Storage Fields

Field	Description
`storageMinGB`	Minimum disk size you can request
`storageMaxGB`	Maximum disk size you can request
`storageCostPerGB`	Hourly cost per GB of storage
`specs.disk`	Included/default disk in GB

When storageMinGB equals storageMaxGB, the disk size is fixed and included with the instance. You do not need to specify disk_gb when deploying.

Get Offer

GET /api/v1/offers/{id}

Returns a single offer by its UUID. Same shape as items in the list response.

Example

curl https://api.deploygpu.ai/api/v1/offers/c0858d33-a3de-4fbd-8fee-98d009c3ec7d

Errors

404 not_found: offer not found.

​List Offers

​Query Parameters

​Response

​Understanding provider.maintenance

​Understanding deploymentType

​Understanding Storage Fields

​Get Offer

​Example

​Errors

List Offers

Query Parameters

Response

Understanding provider.maintenance

Understanding deploymentType

Understanding Storage Fields

Get Offer

Example

Errors