Hugging Face adds Hyperbolic, Nebius AI Studio, and Novita as Serverless Inference Providers

Hugging Face has added Hyperbolic, Nebius AI Studio, and Novita to its ecosystem of serverless inference providers. This expansion increases the variety of models available for serverless inference directly on the Hub's model pages and through official client SDKs.

Expanded Model Availability

The integration of Hyperbolic, Nebius AI Studio, and Novita enables access to several high-demand models, including DeepSeek-R1 and FLUX.1. These new partners join existing providers such as Together AI, Sambanova, Replicate, fal, and Fireworks.ai.

Integration and Usage Workflow

Users can access these inference providers through two primary interfaces: the Hugging Face website UI and the official client SDKs.

Website UI Configuration

In the user account settings, Hugging Face allows users to:

  • Manage API Keys: Users can set custom API keys for specific providers. If no custom key is provided, requests are routed through Hugging Face.
  • Order Preferences: Users can prioritize providers to determine the order in which they appear in model page widgets and code snippets.

Client SDK Implementation

Serverless inference is supported via the huggingface_hub Python library and the @huggingface/inference JavaScript library.

In Python, users can specify the provider argument within the InferenceClient to switch between providers (e.g., provider="hyperbolic" or provider="nebius") while keeping the rest of the code identical. Official support for this functionality is scheduled for release in version v0.29.0 of huggingface_hub.

In JavaScript, the provider parameter is passed within the chatCompletion method of the HfInference client to specify the backend provider (e.g., provider: "novita").

Billing and Authentication Modes

There are two distinct modes for executing inference calls, which determine how billing is handled:

  1. Custom Key Mode: Calls are sent directly to the inference provider using the provider's own API key. Billing is handled entirely by the corresponding provider's account.
  2. Routed Mode: Calls are authenticated via the Hugging Face Hub. Charges are applied directly to the Hugging Face account at standard provider API rates with no additional markup from Hugging Face.

Subscription Benefits

Users on the Hugging Face PRO plan receive $2 worth of Inference credits every month, which can be used across different providers. Signed-in free users are provided with a small quota of free inference.

Sources