# HF Inference API Error on Image Classification Models

**URL:** <https://discuss.huggingface.co/t/hf-inference-api-error-on-image-classification-models/173325>\
**Category:** Inference Endpoints on the Hub\
**Created:** [February 11, 2026, 6:14am UTC](https://discuss.huggingface.co/t/hf-inference-api-error-on-image-classification-models/173325 "2026-02-11T06:14:46Z")\
**Posts on this page:** 1\
**Showing post:** 2

<div class="post-metadata">

**Author:** ![John6666](https://sea2.discourse-cdn.com/hellohellohello/user_avatar/discuss.huggingface.co/john6666/32/27664_2.png) [@John6666](https://discuss.huggingface.co/u/John6666)\
**Post date:** [February 11, 2026, 10:00am UTC](https://discuss.huggingface.co/t/hf-inference-api-error-on-image-classification-models/173325/2 "2026-02-11T10:00:55Z")

</div>

I was able to reproduce that error both from the GUI and from the client, but I don’t know whether these models were actually intentionally deprecated or if they’ll be brought back.

* * *

## What that **410 “model is deprecated”** error usually means

Hugging Face now has **two related (but different) “410 Gone” situations**:

1. **Legacy Inference API shutdown (old domain)**  
If something is still calling `https://api-inference.huggingface.co/...`, Hugging Face returns **410** and tells you to move to the router (`router.huggingface.co`). This is widely reported across libraries and integrations. ([Hugging Face Forums](https://discuss.huggingface.co/t/inference-api-returns-404-not-found-for-all-models/169854 "Inference API returns 404 Not Found for all models"))

2. **Model not served by a specific provider anymore (your case)**  
You are already calling the router with the **HF Inference provider** :  
`https://router.huggingface.co/hf-inference/models/<MODEL_ID>`

The message:

> “The requested model is deprecated and no longer supported by provider hf-inference”

means: **HF Inference (the provider) has stopped serving that model** (or stopped serving it _in the way you’re requesting_), even if the model repo still exists on the Hub.

This matches how the “new” router world works: **a provider serves a catalog, not “every model on the Hub.”** Users forcing `hf-inference` will get errors for models outside that catalog (404/410 are common patterns). ([Hugging Face](https://huggingface.co/docs/inference-providers/providers/hf-inference "HF Inference"))

---

_[View the full topic](https://discuss.huggingface.co/t/hf-inference-api-error-on-image-classification-models/173325)._