Update agent_service.py

agent_service.py

@@ -314,14 +314,14 @@ class AgentService:
             client = AsyncInferenceClient(token=self.hf_token)
             
             # Call HF API with chat completion and native tools
-            # Qwen2.5-72B-Instruct: Best for Vietnamese - state-of-the-art performance
+            # gpt-oss-20B: 21B MoE (3.6B active), fast + good quality, Apache 2.0
             response = await client.chat_completion(
                 messages=messages,
-                model="Qwen/Qwen2.5-72B-Instruct",  # Best for Vietnamese + tool calling
-                max_tokens=1024,  # Increased to prevent truncation
+                model="openai/gpt-oss-20B",  # Balance speed & quality
+                max_tokens=1024,
                 temperature=0.7,
                 tools=tools,
-                tool_choice="auto"  # Let model decide when to use tools
+                tool_choice="auto"
             )
             
             # Check if the model made tool calls