diff --git a/litellm/utils.py b/litellm/utils.py
index 17088f0475b..c2dbbda2b68 100644
--- a/litellm/utils.py
+++ b/litellm/utils.py
@@ -8038,7 +8038,7 @@ def validate_chat_completion_user_messages(messages: list[AllMessageValues]):
 
 def validate_chat_completion_tool_choice(
     tool_choice: dict | str | None,
-    model: str,
+    model: str = "",
 ) -> dict | str | None:
     """
     Confirm the tool choice is passed in the OpenAI format.
diff --git a/tests/litellm_utils_tests/test_validate_tool_choice.py b/tests/litellm_utils_tests/test_validate_tool_choice.py
index b4272af7b90..a9dacf9fa15 100644
--- a/tests/litellm_utils_tests/test_validate_tool_choice.py
+++ b/tests/litellm_utils_tests/test_validate_tool_choice.py
@@ -64,3 +64,11 @@ def test_validate_tool_choice_invalid_type_is_a_400(tool_choice):
     ) as exc_info:
         validate_chat_completion_tool_choice(tool_choice, model=MODEL)
     assert exc_info.value.status_code == 400
+
+
+def test_validate_tool_choice_without_model_is_still_a_400():
+    """Callers that predate the model argument keep getting a 400, with an empty model on the error."""
+    with pytest.raises(litellm.BadRequestError, match="Invalid tool choice") as exc_info:
+        validate_chat_completion_tool_choice({"type": "bogus"})
+    assert exc_info.value.status_code == 400
+    assert exc_info.value.model == ""