Change default Groq model to whisper-large-v3 (#9)

Both models share identical free tier limits (20 RPM, 2K RPD, 8hrs audio/day),
so there is no reason to default to the faster but less accurate turbo variant.
whisper-large-v3 provides better accuracy with negligible latency difference
for short dictation clips.

Author: PatelUtkarsh

AudioType/Core/GroqEngine.swift

@@ -9,7 +9,7 @@ enum GroqModel: String, CaseIterable {
   var displayName: String {
     switch self {
     case .whisperLargeV3Turbo: return "Whisper Large V3 Turbo (faster)"
-    case .whisperLargeV3: return "Whisper Large V3 (most accurate)"
+    case .whisperLargeV3: return "Whisper Large V3 (best, default)"
     }
   }
 
@@ -20,7 +20,7 @@ enum GroqModel: String, CaseIterable {
       {
         return model
       }
-      return .whisperLargeV3Turbo
+      return .whisperLargeV3
     }
     set {
       UserDefaults.standard.set(newValue.rawValue, forKey: "groqModel")