diff --git a/src/microbots/llm/ollama_local.py b/src/microbots/llm/ollama_local.py
index c775631..0ba66d9 100644
--- a/src/microbots/llm/ollama_local.py
+++ b/src/microbots/llm/ollama_local.py
@@ -91,7 +91,8 @@ def _send_request_to_local_model(self, messages):
         headers = {
             "Content-Type": "application/json"
         }
-        response = requests.post(server, json=payload, headers=headers)
+        # Set timeout: 30 seconds connect, 600 seconds read to handle model cold start
+        response = requests.post(server, json=payload, headers=headers, timeout=(30, 600))
         logger.debug(f"\nResponse Code: {response.status_code}\nResponse Text:\n{response.text}\n---")
         if response.status_code == 200:
             response_json = response.json()
diff --git a/test/llm/conftest.py b/test/llm/conftest.py
index 2e75312..a402bfc 100644
--- a/test/llm/conftest.py
+++ b/test/llm/conftest.py
@@ -226,8 +226,48 @@ def ollama_env_config(ollama_model_name, ollama_model_port):
     }
 
 
+@pytest.fixture(scope="session")
+def ollama_model_warmed_up(ollama_server, ollama_env_config):
+    """
+    Warm up the Ollama model by making an initial request.
+
+    This fixture ensures the model is loaded into memory before tests run,
+    which is especially important on CPU-only CI runners where cold starts
+    can cause connection timeouts.
+    """
+    model_name = ollama_env_config["LOCAL_MODEL_NAME"]
+    model_port = ollama_env_config["LOCAL_MODEL_PORT"]
+
+    print(f"\nWarming up Ollama model: {model_name}...")
+
+    try:
+        # Send a simple request to load the model into memory
+        response = requests.post(
+            f"http://localhost:{model_port}/api/generate",
+            json={
+                "model": model_name,
+                "prompt": "hi",
+                "stream": False
+            },
+            headers={"Content-Type": "application/json"},
+            timeout=(30, 600)  # 30s connect, 600s read for model loading
+        )
+
+        if response.status_code == 200:
+            print(f"Model {model_name} warmed up successfully!")
+        else:
+            print(f"Warning: Model warm-up returned status {response.status_code}")
+
+    except requests.exceptions.Timeout:
+        print(f"Warning: Model warm-up timed out. Tests may experience slow first responses.")
+    except Exception as e:
+        print(f"Warning: Model warm-up failed: {e}. Tests may experience slow first responses.")
+
+    return True
+
+
 @pytest.fixture
-def ollama_local_ready(ollama_server, ollama_env_config, monkeypatch):
+def ollama_local_ready(ollama_server, ollama_env_config, ollama_model_warmed_up, monkeypatch):
     """
     Complete setup fixture that ensures Ollama is ready for testing.