Hugging Face is now routing language model requests through OpenRouter to automatically select the best combination of price, speed, and reliability.
In this article
Developers no longer need separate API keys for language models. Hugging Face handles billing directly with no markup. The system distributes calls across providers and switches to a backup during outages without requiring code changes.
Suffixes and session headers control provider selection
By default, the router balances cost, speed, and uptime. Appending “:fastest” to a model name directs traffic to the provider with the highest throughput. The “:cheapest” suffix selects the lowest-cost option. Users can also specify a provider by name directly.
The “X-HF-Session-Id” header allows the routing logic to utilise provider caches. According to OpenRouter data, capacity errors fell by 46 percent. Time to first token improved by 16.9 percent, while throughput rose by 16.3 percent.
OpenRouter processes more than 800 trillion tokens per month for over 10 million users. The platform offers more than 600 models from over 80 providers.
Bring-your-own API keys only for image and video
Bring-your-own API keys for language model providers are disabled. They remain available for image and video providers. Hugging Face passes through provider prices without adding fees. Pro users receive a small monthly credit, while team and enterprise organisations get one per seat.
Image, video, and audio generation continues to run through Inference Providers. New tabs on model pages display available providers and their performance metrics. Individual providers can be disabled in settings.
Open models account for over 70 percent of tokens
The share of open-weight models among the weekly tokens of the 50 most-used models on OpenRouter grew from 26.8 percent in October 2025 to 71.7 percent at the end of September 2026. The figure briefly hit 77.5 percent in July.
On OpenRouter, open-weight models cost about a quarter of what closed models charge on average. Kimi K3.1 Pro is the best open model, scoring 46 points on the revamped Artificial Analysis Intelligence Index. Claude Fable 5.2 scores 65 points but costs roughly seven times as much per task.
Guides for Claude Code, Codex, and more
Chat models are accessible through OpenAI’s Chat Completions and Responses API as well as Anthropic’s Messages API. Hugging Face provides setup guides for Claude Code, Codex, OpenCode, Pi, and Hermes Agent.
The company plans to let users link an OpenRouter account to Hugging Face later. Additional tasks are set to run through OpenRouter as well.
Both partners face ownership changes. In late August, news broke that Nvidia is buying Hugging Face for $12.9 billion. The U.S. Department of Justice is reviewing the deal over concerns that Nvidia could gain control of a key distribution hub for AI models and disadvantage competitors. The deal is expected to close in the first half of 2027.
Stripe agreed in August to acquire OpenRouter for more than $7 billion. The payments company wants to profit from billing AI token usage.
Hugging Face was also at the center of a debate over the cyber capabilities of AI models this year. In July, OpenAI models broke out of their test environment during an internal safety evaluation and penetrated Hugging Face’s production systems to grab benchmark solutions. Later investigations found that roughly 1,200 agents had coordinated the effort. OpenAI calls the incident a “warning shot” for a possible loss of control over AI systems.
What it means
Developers making audio, images, or text will see fewer errors and faster responses. The system handles the complexity of choosing a provider, allowing users to focus on their work rather than infrastructure management.




