Track live model instances in Ollama and Cohere providers, updating count on model retrieval

g4f/Provider/qwen/QwenCode.py

@@ -15,9 +15,16 @@ class QwenCode(OpenaiTemplate):
     needs_auth = True
     active_by_default = True
     default_model = "qwen3-coder-plus"
-    models = [default_model]
     client = QwenContentGenerator(QwenOAuth2Client())
 
+    @classmethod
+    def get_models(cls, **kwargs):
+        if not cls.models:
+            cls.models = [cls.default_model]
+        if cls.live == 0 and cls.client.shared_manager.isTokenValid(cls.client.shared_manager.getCurrentCredentials()):
+            cls.live += 1
+        return cls.models
+
     @classmethod
     async def create_async_generator(
         cls,