diff --git a/pybossa/api/__init__.py b/pybossa/api/__init__.py
index 9b2f104a7..07d612ff5 100644
--- a/pybossa/api/__init__.py
+++ b/pybossa/api/__init__.py
@@ -879,7 +879,7 @@ def large_language_model(model_name):
     model_endpoint = endpoints.get(model_name.lower())
 
     if not model_endpoint:
-        return abort(403, f'Model with name {model_name} does not exist.')
+        return abort(400, f'{model_name} LLM is unsupported on this platform.')
 
     proxies = current_app.config.get('PROXIES')
     cert = current_app.config.get('CA_CERT', False)
@@ -899,7 +899,7 @@ def large_language_model(model_name):
         if isinstance(prompts, list):
             prompts = prompts[0]  # Batch request temporarily NOT supported
         if not isinstance(prompts, str):
-            return abort(400, f'prompts should be a string a list of strings')
+            return abort(400, f'prompts should be a string or a list of strings')
         data = {
             "instances": [
                 {
diff --git a/test/test_api/__init__.py b/test/test_api/__init__.py
index 22f78007a..d939b2d6e 100644
--- a/test/test_api/__init__.py
+++ b/test/test_api/__init__.py
@@ -125,8 +125,8 @@ def test_invalid_model_name(self, mock_post):
             "prompts": "Identify the company name: Microsoft will release Windows 20 next year."
         }):
             response = large_language_model('invalid-model')
-            self.assertEqual(response.status_code, 403)
-            self.assertIn('Model with name', response.json.get('exception_msg'))
+            self.assertEqual(response.status_code, 400)
+            self.assertIn('LLM is unsupported', response.json.get('exception_msg'))
 
     @patch('requests.post')
     def test_invalid_json(self, mock_post):