Get a GPU type
Returns a single GPU type with pricing. Availability details are included only when requested with include=AVAILABILITY.
Authorizations
Runpod API key authentication. Generate an API key in the Runpod console and send it in the Authorization header as Bearer <api_key>. Keys are scoped to the permissions granted when created; requests may return 403 when a valid key lacks access to the requested resource or action.
Path Parameters
Query Parameters
Comma-separated optional expansions. Supported value today: AVAILABILITY. This may expand with more include values in the future.
1Catalog include expansion. Only AVAILABILITY is supported today; additional include values may be added in the future.
AVAILABILITY Comma-separated availability product contexts. Supported values: POD, CLUSTER, SERVERLESS. Valid only with include=AVAILABILITY. Upstream default when omitted: POD.
Catalog product availability context.
POD, CLUSTER, SERVERLESS GPU count for availability and lowest-price calculations. Valid only with include=AVAILABILITY. Defaults to 1.
x >= 1Cloud type for availability and lowest-price calculations. Valid only with include=AVAILABILITY. Supported values: SECURE, COMMUNITY. Upstream default when omitted: SECURE. GPU availability cloud filter.
SECURE, COMMUNITY Comma-separated CUDA versions to scope availability and lowest-price calculations to, matched exactly. Format: major.minor, e.g. 12.8 — a bare major is rejected here because it identifies no version. Valid only with include=AVAILABILITY (400 otherwise) and mutually exclusive with minCudaVersion (400 if both are sent); a malformed entry is a 422. Also narrows the returned cudaVersions array; omit it to enumerate every version offered.
^\d+\.\d+$Lowest acceptable CUDA version to scope availability and lowest-price calculations to, compared numerically. Format: integer major or major.minor, e.g. 12 or 12.1 — a bare major is accepted here and means any release of that major. Valid only with include=AVAILABILITY (400 otherwise) and mutually exclusive with cudaVersions (400 if both are sent); a malformed value is a 422. Use this for an open-ended floor and cudaVersions for an exact set.
^\d+(\.\d+)?$Response
OK
Individual GPU type identifier (use for pod creation)
"NVIDIA GeForce RTX 4090"
"RTX 4090"
Serverless GPU pool ID (use for serverless endpoint creation). Null if GPU is not in a serverless pool.
"ADA_24"
Canonical GPU hardware manufacturer.
NVIDIA, AMD, UNKNOWN VRAM in GB
24
Available on secure cloud
true
Available on community cloud
true
Overall GPU availability. Present only when requested with include=AVAILABILITY.
NONE, LOW, MEDIUM, HIGH Per-datacenter GPU availability. Present only when requested with include=AVAILABILITY.
CUDA versions offered by machines with this GPU type, each tagged
with current capacity. Present only when requested with
include=AVAILABILITY, and scoped by the same filters as
availability (count, cloud, product, and whichever of
cudaVersions / minCudaVersion was supplied).
Machines that report no CUDA version are skipped, so this property
is absent entirely for a GPU type with none — AMD, for instance.
Treat a missing cudaVersions the same as an empty one. A version
absent from a populated list is not offered for this GPU type.