fix models

agent.py

@@ -57,8 +57,8 @@ def smart_invoke(msgs, use_tools=False, start_tier=0):
     Retries next tier if a 429 (rate limit), 402 (credits), or 404 (model found) error occurs.
     """
     
-    # Adaptive Gemini names to try if 3.1 flash is 404
-    gemini_alternatives = ["gemini-2.0-flash", "gemini-3.1-flash-lite", "gemini-3.1-pro"]
+    # Adaptive Gemini names verified via list_models (REST API)
+    gemini_alternatives = ["gemini-2.5-flash", "gemini-2.0-flash", "gemini-flash-latest", "gemini-pro-latest"]
     
     tiers_config = [
         {"name": "OpenRouter", "key": "OPENROUTER_API_KEY", "provider": "openai", "model_name": "meta-llama/llama-3.3-70b-instruct", "base_url": "https://openrouter.ai/api/v1"},
@@ -179,11 +179,10 @@ def get_vision_models():
     """Returns a list of vision models to try, in order of preference."""
     configs = [
         {"name": "OpenRouter-Gemini-2.0", "key": "OPENROUTER_API_KEY", "provider": "openai", "model_name": "google/gemini-2.0-flash-001", "base_url": "https://openrouter.ai/api/v1"},
-        {"name": "Google-Gemini-2.0-Exp", "key": "GOOGLE_API_KEY", "provider": "google", "model_name": "gemini-2.0-flash-exp"},
-        {"name": "Google-Gemini-1.5-Latest", "key": "GOOGLE_API_KEY", "provider": "google", "model_name": "gemini-1.5-flash-latest"},
+        {"name": "Google-Gemini-2.0-Flash", "key": "GOOGLE_API_KEY", "provider": "google", "model_name": "gemini-2.0-flash"},
+        {"name": "Google-Gemini-Flash-Latest", "key": "GOOGLE_API_KEY", "provider": "google", "model_name": "gemini-flash-latest"},
         {"name": "NVIDIA-Vision-Llama-11b", "key": "NVIDIA_API_KEY", "provider": "openai", "model_name": "meta/llama-3.2-11b-vision-instruct", "base_url": "https://integrate.api.nvidia.com/v1"},
         {"name": "NVIDIA-Vision-Llama-90b", "key": "NVIDIA_API_KEY", "provider": "openai", "model_name": "meta/llama-3.2-90b-vision-instruct", "base_url": "https://integrate.api.nvidia.com/v1"},
-        {"name": "Groq-Vision", "key": "GROQ_API_KEY", "provider": "groq", "model_name": "llama-3.2-90b-vision-preview"},
     ]
     models = []
     for cfg in configs:
@@ -356,7 +355,10 @@ def analyze_video(video_path: str, question: str) -> str:
         
         return f"Video Summary based on extracted frames and audio:\n{video_context}"
     except Exception as e:
-        return f"Error analyzing video: {str(e)}"
+        err_msg = str(e)
+        if "No address associated with hostname" in err_msg or "Failed to resolve" in err_msg:
+            return f"Error: The environment cannot access the internet (DNS failure). Please use 'web_search' or 'wiki_search' to find information about this video content instead of trying to download it."
+        return f"Error analyzing video: {err_msg}"
     finally:
         if downloaded_video and os.path.exists(downloaded_video):
             try:
@@ -520,7 +522,12 @@ def answer_message(state: AgentState) -> AgentState:
             time.sleep(4)
             
         print(f"--- ReAct Step {step + 1} ---")
-        ai_msg, current_tier = smart_invoke(messages, use_tools=True, start_tier=current_tier)
+        
+        # Max history truncation to avoid 413 Request Too Large errors
+        # Keep SystemMessage, first HumanMessage, and the last 6 messages
+        safe_messages = messages[:2] + messages[-6:] if len(messages) > 10 else messages
+        
+        ai_msg, current_tier = smart_invoke(safe_messages, use_tools=True, start_tier=current_tier)
         messages.append(ai_msg)
 
         # Check if the model requested tools