Use Vertex API for accessing Google models!

Using Gemini 2.5 Pro on t3.chat is really really slow. Please use the Vertex API instead of Google AI studio for faster speeds as we can see on OpenRouter that the latency is way shorter there.