Hugging Face is teaming up with OpenRouter to make open-weight models easier to run. Every major open-weight release already lands on the Hugging Face Hub, and OpenRouter serves those models at scale, routing each request across more than 80 providers. From today, Hugging Face's Inference Providers use OpenRouter's router instead of its own: language-model requests go through OpenRouter, which picks the best provider for each call based on live price, speed and reliability, and retries with another provider if the first one fails. Code stays the same, and the same Hugging Face token works across the Hub.
The shift comes as open-weight models surge on OpenRouter. They grew from 26.8% of the tokens served in October 2025 to 71.7% at the end of September 2026, peaking at 77.5% in July, with weekly open-weight volume up more than 60-fold over the same period. OpenRouter now serves more than 800 trillion tokens a month to over 10 million users, across 600-plus models from more than 80 providers.
Open-weight models also cost about four times less on average: $0.28 per million input tokens and $1.62 per million output tokens, against $1.27 and $6.36 for closed models. The best open-weight models still trail the best proprietary ones on the hardest benchmarks — Kimi K3.1 Pro scores 46 on the Artificial Analysis Intelligence Index against 65 for Claude Fable 5.2 — but at about a seventh of the cost per task, $3.68 against $26.54.
OpenRouter's rebuilt routing cut capacity errors by 46% and improved time to first token by 16.9% and throughput by 16.3% in its measurements. Pricing does not change, and Hugging Face adds no markup on top of provider prices.
One breaking change comes with the move: requests to language models are now billed to the Hugging Face account, so developers can no longer use their own API key for a provider with these models. Providers can still be picked by suffix or switched off in settings. Image, video and audio generation are unaffected and keep working as before.





