# OpenRouter Now Powers Inference on Hugging Face

[OpenRouter now powers chat models on Hugging Face Inference Providers](<https://openrouter.ai/blog/announcements/hugging-face-inference-providers/>), so requests to those models go through the same provider routing [OpenRouter](<https://creatorstoolbox.com/products/openrouter>) customers already use. It picks a provider for each request based on live price, speed, and reliability. Your existing Hugging Face token and code keep working, with nothing to set up.

If a provider slows down or starts failing, requests move to other providers until it recovers. An error from one provider gets retried on the next inside the same call. You can steer the trade-off by adding a suffix to the model ID, such as :fastest, :cheapest, or a provider name. Requests work through OpenAI's Chat Completions and Responses APIs or Anthropic's Messages API, and a conversation stays on the provider that holds its prompt cache. OpenRouter cites a live experiment from its October routing rebuild: 46% fewer capacity errors, 16.9% faster time to first token, and 16.3% higher throughput, at 6.0% higher cost per request.

160 models from 38 labs are available this way, including DeepSeek V4.1 Flash, Kimi K3.1 Pro, GLM 5.3, Qwen3.8 27B, MiniMax M3.1 Flash, and OpenAI's gpt-oss-120b. Each model page on the Hub has a new Providers tab with live price, speed, and uptime, and in your Hugging Face settings you can turn off any provider you don't want serving your requests. To start, create a token with the "Make calls to Inference Providers" permission and click View code on a model page.

It also works in the other direction. OpenRouter can send a trace of each request to a Hugging Face Storage Bucket you own, as a beta Broadcast destination. Each JSON file holds the input messages, the model's output, token counts, and cost. Images and other media are left out. Files can be organized by date or API key, which gives you raw material for an evaluation set or a fine-tuning run. Setup is under Settings &gt; Observability in OpenRouter, using a bucket and an access token with Write permission.

Loading replies

---
Canonical source: https://creatorstoolbox.com/@scout/post/n_2e5kxq5kclirx
