diff --git a/AppImage/config/verified_ai_models.json b/AppImage/config/verified_ai_models.json
index fb81221d..2e0c716d 100644
--- a/AppImage/config/verified_ai_models.json
+++ b/AppImage/config/verified_ai_models.json
@@ -18,12 +18,12 @@
   "gemini": {
     "models": [
       "gemini-2.5-flash",
-      "gemini-2.5-pro",
-      "gemini-1.5-flash"
+      "gemini-2.5-flash-lite",
+      "gemini-2.5-pro"
     ],
-    "recommended": "gemini-1.5-flash",
-    "_note": "gemini-2.5-* may have availability issues. gemini-1.5-flash is more stable.",
-    "_deprecated": ["gemini-2.0-flash", "gemini-2.0-flash-lite", "gemini-1.0-pro", "gemini-pro"]
+    "recommended": "gemini-2.5-flash",
+    "_note": "gemini-2.5-flash-lite is cheaper but may struggle with complex prompts. Use with simple/custom prompts.",
+    "_deprecated": ["gemini-2.0-flash", "gemini-2.0-flash-lite", "gemini-1.5-flash", "gemini-1.0-pro", "gemini-pro"]
   },
   
   "openai": {
diff --git a/AppImage/scripts/ai_providers/base.py b/AppImage/scripts/ai_providers/base.py
index 8790f188..1f6e1a60 100644
--- a/AppImage/scripts/ai_providers/base.py
+++ b/AppImage/scripts/ai_providers/base.py
@@ -65,7 +65,7 @@ class AIProvider(ABC):
             response = self.generate(
                 system_prompt="You are a test assistant. Respond with exactly: CONNECTION_OK",
                 user_message="Test connection",
-                max_tokens=20
+                max_tokens=50  # Some providers (Gemini) need more tokens to return any content
             )
             if response:
                 # Check if response contains our expected text
diff --git a/AppImage/scripts/ai_providers/gemini_provider.py b/AppImage/scripts/ai_providers/gemini_provider.py
index 62a3e1c7..c63b9ea8 100644
--- a/AppImage/scripts/ai_providers/gemini_provider.py
+++ b/AppImage/scripts/ai_providers/gemini_provider.py
@@ -166,8 +166,15 @@ class GeminiProvider(AIProvider):
                 if text:
                     return text
             
-            # No text content - provide detailed error
-            raise AIProviderError(f"No text in response (finishReason: {finish_reason})")
+            # No text content - check if it's a known issue
+            if finish_reason == 'MAX_TOKENS':
+                # MAX_TOKENS with no content means the prompt itself was too long
+                raise AIProviderError("Prompt too long - no space for response. Try a shorter system prompt or message.")
+            elif finish_reason == 'STOP':
+                # Normal stop but no content - unusual
+                raise AIProviderError("Model returned empty response")
+            else:
+                raise AIProviderError(f"No text in response (finishReason: {finish_reason})")
         except AIProviderError:
             raise
         except (KeyError, IndexError) as e: