If you want to provide inference on OpenRouter—whether for existing models or new ones—please use our dedicated provider application instead.