NexusComputeBETA
Compute for AI builders

GPU inference compute without the cloud lock-in

Explore provider-run GPUs for LLM inference, inspect the details that matter, and compare available capacity before you begin a marketplace connection.

Browse GPU compute Read the API docs

Looking to contribute hardware instead? Learn about earning from idle compute.

Network snapshot

These counters come from the marketplace and can change as providers and jobs change.

—
Active providers
—
Available listings
—
Jobs completed
—
Tokens processed

Choose by workload

Find the right GPU inference capacity

Start with the model and hardware requirements of your workload, then compare individual listings rather than choosing an opaque bundle.

Model fit
Filter for listings serving the AI model your application needs.
GPU and VRAM
Review GPU model, count, and VRAM before sending inference traffic.
Price and access
Compare on-demand, spot, and reserved options when a provider offers them.

Available compute to explore

Live listing previews; availability and pricing can change.

See all listings

From search to request

A clear path to your first inference call

  1. 01

    Browse

    Filter the marketplace by model, availability, provider, and instance type.

  2. 02

    Inspect

    Open a listing to review the provider, hardware, pricing, and connection details.

  3. 03

    Connect

    Use the marketplace connection flow and review the REST API resources relevant to your application.

  4. 04

    Plan

    Use the listing details and API reference to plan how your application will work with the marketplace.

What you can use today

  • A marketplace for provider listings and model availability
  • A marketplace for comparing live provider listings before you connect
  • Displayed provider rates and a $0.10 per 1,000-token platform fee
  • Public REST API documentation for listings, endpoints, connections, jobs, billing, and keys
Compare pricingBrowse endpointsOpen API reference

GPU inference questions

The practical details to check before you connect an application.

How do I access GPU inference compute?
Browse available NexusCompute listings and review the model, GPU, VRAM, location, availability, and displayed price before using the marketplace connection flow.
How is GPU inference priced?
NexusCompute displays a provider rate plus the platform fee per 1,000 tokens. The marketplace also labels spot and reserved options when a listing offers them.
Can I use the NexusCompute API in my application?
The public REST API documentation explains authentication, listings, endpoints, connections, jobs, billing, and API keys.

Ready to compare GPU compute?

Browse the live marketplace or read the API reference before you connect an application.

Browse compute Read API docs