My Gemini API Query is taking forever to respond, how do I make it faster?

Gemini API
,
1

I am currently designing a prototype chatbot for a school project, but I was confused as to why the API call was lagging so much. It takes a minimum of 2 minutes to respond. Here is my code below.

import Foundation
import SwiftUI
import GoogleGenerativeAI

enum ChatRole {
    case user
    case model
}

struct ChatMessage: Identifiable, Equatable {
    let id = UUID().uuidString
    var role: ChatRole
    var message: String
}

class ChatService: ObservableObject {
    @Published private var chat: Chat?
    @Published private(set) var messages = [ChatMessage]()
    @Published private(set) var loadingResponse = false
    
    func sendMessage(_ message: String) {
        loadingResponse = true
        if (chat == nil) {
     let history: [ModelContent] = messages.map { ModelContent(role: $0.role == .user ? "user" : "model", parts: $0.message)}
            chat = GenerativeModel(name: "gemini-2.0-flash-lite-preview-02-05", apiKey: APIKey.default, systemInstruction: "You are a financial AI assistant. You are called Assist, and is supposed to do nothing other than answer questions.").startChat(history: history)
 }
 
 // MARK: Add user's message to the list
 messages.append(.init(role: .user, message: message))
 
 Task {
     do {
         let response = try await chat?.sendMessage(message)
         
         loadingResponse = false
         
         guard let text = response?.text else {
             messages.append(.init(role: .model, message: "Something went wrong, please try again."))
             return
         }
         
         messages.append(.init(role: .model, message: text))
     }
     catch {
         loadingResponse = false
         messages.append(.init(role: .model, message: "Something went wrong, please try again."))
     }
 }
}
}

I’ve tried to give it specific instructions, changed models, but nothing is working. I’ve also tried it on different devices and WiFi networks for no avail. Is it my code or is it a API problem?