Inferencing · Provider Landscape

Inference Provider Explorer

Explore providers integrated with Hugging Face Inference Providers by workload. Use this as a first-pass capability map, then verify live model-specific metrics and features on Hugging Face.

LLMsVLMsEmbeddings ImageVideoSpeech

Filter by workload

Select a workload to see providers documented by Hugging Face as supporting that inference category.

Reference snapshot: September 2026

Important: this explorer does not hard-code live latency, throughput, pricing, tool-calling, or structured-output values. Those signals can change by model and provider. Use Hugging Face's live provider/model comparison for current model-specific data.

How to use this explorer

1. Narrow the provider landscapeStart with the workload your application actually needs.

2. Check the exact modelProvider availability is model-specific. A provider supporting LLM chat does not imply it serves every LLM.

3. Verify capabilitiesCheck tool calling, structured outputs, context, streaming, and other endpoint requirements.

4. Compare live performanceReview current first-token latency, throughput, pricing, and other model/provider signals where available.

Primary sources

This project uses Hugging Face documentation as the primary reference for the provider capability snapshot.

Inference Providers documentationOpen official documentation

Live supported models & provider metricsOpen live provider comparison

Hub API for provider mappingsOpen Hub API documentation

Pricing & billingOpen pricing documentation