I’ve been using the Gemini API via Google AI to work with Gemini 1.5 Pro, but I’ve noticed that the responses can sometimes be slow (regardless of the number of tokens). I’m planning to set up a chatbot for our website and recently learned about using the Gemini API through Vertex AI. For a production environment, would it be better to use Vertex AI instead? Is it generally more stable than accessing the Gemini API directly through Google AI?Also, what are the other benefits of using Vertex AI over Google AI?