Public AI added as a Hugging Face Inference Provider via Public AI Utility (Apertus-70B-Instruct-2509)

AI Impact Summary

Public AI is now a supported Inference Provider on the Hugging Face Hub, enabling serverless inference directly from model pages and through HF client SDKs. There are two billing modes: Custom key routing (requests go directly to the provider) and Routed by HF (HF bills, provider charges pass through HF), which changes credential handling and cost flow for applications. The integration highlights models like swiss-ai/Apertus-70B-Instruct-2509 and relies on a vLLM-powered backend exposing OpenAI-compatible APIs, with global load balancing and a free credit program for PRO users.