diff --git a/litellm/litellm_core_utils/prompt_templates/factory.py b/litellm/litellm_core_utils/prompt_templates/factory.py
index 90a60a9e46c..b60a735170c 100644
--- a/litellm/litellm_core_utils/prompt_templates/factory.py
+++ b/litellm/litellm_core_utils/prompt_templates/factory.py
@@ -1144,9 +1144,7 @@ def _gemini_tool_call_invoke_helper(
     if (
         isinstance(arguments, str) and len(arguments) == 0
     ):  # pass empty dict, if arguments is empty string - prevents call from failing
-        arguments_dict = {
-            "type": "object",
-        }
+        arguments_dict = {}
     else:
         arguments_dict = json.loads(arguments)
     function_call: Final = VertexFunctionCall(
diff --git a/tests/unit/litellm_core_utils/prompt_templates/test_litellm_core_utils_prompt_templates_factory.py b/tests/unit/litellm_core_utils/prompt_templates/test_litellm_core_utils_prompt_templates_factory.py
index 8d2e6b9fd0c..ca488643ac2 100644
--- a/tests/unit/litellm_core_utils/prompt_templates/test_litellm_core_utils_prompt_templates_factory.py
+++ b/tests/unit/litellm_core_utils/prompt_templates/test_litellm_core_utils_prompt_templates_factory.py
@@ -593,9 +593,33 @@ def test_vertex_ai_transform_empty_function_call_arguments():
     }
     result: VertexFunctionCall = _gemini_tool_call_invoke_helper(function_call)
     print(result)
-    assert result["args"] == {
-        "type": "object",
-    }
+    assert result["args"] == {}
+
+
+def test_gemini_missing_arguments_maps_to_empty_object():
+    """When `arguments` key is absent, the helper should yield an empty object (see #43156)."""
+    from litellm.litellm_core_utils.prompt_templates.factory import (
+        VertexFunctionCall,
+        _gemini_tool_call_invoke_helper,
+    )
+
+    result: VertexFunctionCall = _gemini_tool_call_invoke_helper(
+        {"name": "no_args"}
+    )
+    assert result["args"] == {}
+
+
+def test_gemini_nonempty_arguments_is_parsed():
+    """Non-empty arguments string should still be JSON-decoded as before."""
+    from litellm.litellm_core_utils.prompt_templates.factory import (
+        VertexFunctionCall,
+        _gemini_tool_call_invoke_helper,
+    )
+
+    result: VertexFunctionCall = _gemini_tool_call_invoke_helper(
+        {"name": "get", "arguments": '{"x": 1}'}
+    )
+    assert result["args"] == {"x": 1}
 
 
 @pytest.mark.asyncio