update for mcp

backend/app/services/llm/llm_factory.py

@@ -16,7 +16,7 @@ from app.shared.model_usage_tracker import get_model_usage_tracker
 # Keep provider-specific behavior explicit so debugging stays straightforward.
 DEFAULT_MODELS = {
     LLMProvider.DEEPSEEK: "deepseek-v4-flash",
-    LLMProvider.QWEN: "qwen3.5-flash",
+    LLMProvider.QWEN: "qwen3.6-flash",
     LLMProvider.QWEN_VL: "qwen3-vl-plus"
 }
 
@@ -101,6 +101,8 @@ class LLMFactory:
             "qwen-max": LLMProvider.QWEN,
             "qwen3.5-flash": LLMProvider.QWEN,
             "qwen3.5-plus": LLMProvider.QWEN,
+            "qwen3.6-flash": LLMProvider.QWEN,
+            "qwen3.6-plus": LLMProvider.QWEN,
             "qwen_vl": LLMProvider.QWEN_VL,
             "qwen-vl": LLMProvider.QWEN_VL,
             "qwen-vl-plus": LLMProvider.QWEN_VL,

backend/app/services/llm/qwen_client.py

@@ -27,6 +27,8 @@ class QwenClient(BaseLLMClient):
         "qwen-long",
         "qwen3.5-flash",
         "qwen3.5-plus",
+        "qwen3.6-flash",
+        "qwen3.6-plus",
         "qwen3-plus",
         "qwen2.5-72b-instruct",
         "qwen2.5-32b-instruct",
@@ -371,7 +373,7 @@ class QwenVLClient(BaseLLMClient):
 
 def create_qwen_client(
     api_key: str,
-    model: str = "qwen3.5-flash",
+    model: str = "qwen3.6-flash",
     base_url: str = "http://6.86.80.4:30080/v1",
     **kwargs
 ) -> QwenClient: