Registry / Tools / Hugging Face Inference Endpoints

Hugging Face Inference Endpoints facts

tool_hugging_face_inference_endpoints
Platform

Hugging Face Inference Endpoints

Managed infrastructure for deploying Hugging Face models as dedicated inference endpoints.

last verified: 2026-05-06pricing checked: 2026-05-06
01

Fact summary

Name
Hugging Face Inference Endpoints
Company / maker
Hugging Face
Tags
Model Hosting, Inference, Deployment
Entity type
Platform
Availability
Public
Pricing model
Usage Based
Pricing unit
compute time
Primary pricing route
Api Usage
Free plan
No
Delivery modes
Web AppAPISDK
Supported platforms
Web
Data state
source backed
02

Official links

Formal source rows with URL, source type, and freshness.
LinkURLsource typechecked
Official websitehuggingface.co/docs/inference-endpoints/en/index official-site2026-05-06
Pricing pagehuggingface.co/docs/inference-endpoints/en/pricing official-pricing-page2026-05-06
Documentationhuggingface.co/docs/inference-endpoints/en/index official-documentation2026-05-06
API docshuggingface.co/docs/inference-endpoints/en/api_reference official-documentation2026-05-06
03

Access and delivery

Supported platforms and delivery modes are intentionally separate.

Supported platforms

Web

Delivery modes

Web AppAPISDK
04

Capabilities

Capability tags

Model HostingModel ApisEvaluation Monitoring
Input types
Structured DataFunction CallTextImageAudioVideo
Output types
Structured JsonTextImageAudioVideoEmbeddingsDocument
05

Pricing summary

Pricing summary

Derived from official pricing routes where available
pricingModel
Usage Based
pricingUnit
compute time
currency
USD
hasFreePlan
No
pricingLastChecked
2026-05-06
06

Pricing routes

Each route is a first-class record with source URL and checked date.
Request a quote
routeType: enterprise-sales

The pricing page includes a request-a-quote path and notes quota requests for unavailable instance types.

No numeric price asserted
Dedicated endpoint compute
routeType: api-usage
primary route

Dedicated endpoints are priced by selected instance type; the pricing page states hourly rates are shown and actual cost is calculated by the minute while deployed endpoints are initializing or running.

No numeric price asserted / compute time
07

Pricing plans

plantrackusage limitfeaturessourcechecked
Usage-based dedicated endpointsapiCosts are calculated per minute from the selected instance hourly rate while endpoints are initializing and running.Dedicated Inference Endpoints, Select instance type to deploy and scale models, CPU, GPU, and accelerator instance hourly pricing, Accessible to Hugging Face accounts with an active subscription and credit card on file2026-05-06
08

Evidence and sources

source URLsource typeused forchecked
huggingface.co/docs/inference-endpoints/en/pricingofficial-pricing-pagepricing-model, pricing-routes2026-05-06
huggingface.co/docs/inference-endpoints/en/api_referenceofficial-documentationapi-access, delivery-modes, capability-claims2026-05-06
huggingface.co/docs/inference-endpoints/indexofficial-documentationidentity, capabilities, delivery-modes2026-05-06
09

Freshness and recent changes

Facts verified
2026-05-06
Identity and core facts
Pricing checked
2026-05-06
Pricing routes and plans
Record updated
2026-05-06
Public record freshness
10

Related pages

One-hop registry pages derived from this record's modeled facts. This section does not infer peer tools.