DeepInfra added as a Hugging Face Inference Provider with DeepSeek V4 Pro support

AI Impact Summary

DeepInfra is now a supported Inference Provider on the Hugging Face Hub, enabling serverless inference directly from model pages and through HF’s client SDKs. There are two routing modes: Custom key (direct to the provider) and Routed by HF (charges billed to Hugging Face), which will influence authentication flows and cost attribution for applications. This expands access to DeepInfra’s catalog (e.g., DeepSeek V4 Pro) and sets the stage for broader task support, impacting deployment patterns and monitoring for multi-provider inference setups.