Skip to main content

Overview

The get_available_gpus action retrieves all available GPU machines on the Hyperbolic platform, including pricing, specifications, and availability information.

Function Signature

def get_available_gpus() -> str

Input Schema

This action takes no input parameters.

Response

Returns a formatted string with details about available GPU options.
output
string
A formatted string containing:
  • Cluster name
  • Node ID
  • GPU model
  • Available GPUs (available/total)
  • Price per GPU per hour in USD

API Endpoint

POST https://api.hyperbolic.xyz/v1/marketplace

Example Response

Available GPU Options:

Cluster: us-east-1
Node ID: node-abc123
GPU Model: NVIDIA H100
Available GPUs: 4/8
Price: $2.50/hour per GPU
----------------------------------------

Cluster: us-west-2
Node ID: node-def456
GPU Model: NVIDIA A100
Available GPUs: 2/4
Price: $1.75/hour per GPU
----------------------------------------

Example Usage

from hyperbolic_agentkit_core.actions import get_available_gpus

# Get all available GPU options
available_gpus = get_available_gpus()
print(available_gpus)

Important Notes

  • Authorization key (HYPERBOLIC_API_KEY) is required for this operation
  • GPU prices are returned in dollars per hour (converted from cents)
  • Only non-reserved instances with available GPUs are displayed
  • If no GPU instances are available, returns: “No available GPU instances found.”
  • Use the cluster_name and node_name from this response when calling rent_compute

Response Details

The function processes the marketplace API response and filters for:
  • Non-reserved instances (reserved: false)
  • Instances with available GPUs (gpus_available > 0)
  • Instances with valid hardware specifications
For each matching instance, the response includes:
  • Cluster: Geographic or logical cluster identifier
  • Node ID: Unique node identifier for use in rent_compute
  • GPU Model: Hardware model (e.g., NVIDIA H100, A100, etc.)
  • Available GPUs: Number of GPUs available out of total
  • Price: Cost per GPU per hour in USD

Build docs developers (and LLMs) love