Loading models…

Fetching the model catalog.

↑↓ Navigate ↵ OpenBrowse all models ↗

Home Models Fusions Rankings Compare Docs Workspace

Pay for requests with

Personal workspace

Pay for requests with

WorkspaceOverview Playground Fusions API Keys Batches

AccountCredits & wallets Activity Earned credits Settings

ResourcesModels Quickstart Pricing

Terms of Service Privacy Policy Acceptable Use Service Disclosures

WORKSPACE

Fusions

Choose your models, tune your routing, and track your savings.

One model ID. Your choice of models.

A Fusion routes each request to a model from your saved lineup. Use its model ID in your coding client or your own app, and adjust the routing here whenever you need to.

To get started, choose or create a Fusion below, configure its models, and copy its model ID. Keep the Fusion enabled and create an API key with credits available on your account.

In a coding client

Use your Fusion in clients such as Cursor, Cline, Zed, Continue, OpenCode, or Aider.

Follow your client’s setup guide to configure GPU Router and its base URL.
Enter your GPU Router API key and use your Fusion’s model ID as the custom model.
Select that model and start a chat. Fusion routes the client’s requests using your settings.

View coding client setup guides →

Through the API

Call your Fusion from a script, backend, or OpenAI-compatible SDK.

Set the base URL to https://gpurouter.io/v1 and authenticate with your GPU Router API key.
Send a Chat Completions request with your Fusion’s ID in the model field and your prompt in messages.
Open Connection details → Call Fusion via API on the Fusion below for a copyable curl example with its model ID filled in.

Keep your API key on your server or in an environment variable.

After your first request, check Activity for the model used, or open Usage & savings on your Fusion.

Your Fusions

Pick your models once. Your Fusion chooses a route for every prompt.