diff --git a/litellm/utils.py b/litellm/utils.py
index 8d55783bf9c..4bb85c46249 100644
--- a/litellm/utils.py
+++ b/litellm/utils.py
@@ -1033,17 +1033,47 @@ def function_setup(  # noqa: PLR0915
             call_type == CallTypes.image_generation.value
             or call_type == CallTypes.aimage_generation.value
         ):
-            messages = args[0] if len(args) > 0 else kwargs["prompt"]
+            if len(args) > 0:
+                messages = args[0]
+            else:
+                prompt_value = kwargs.get("prompt")
+                if prompt_value is None:
+                    raise BadRequestError(
+                        message="Missing required parameter: 'prompt' for image generation.",
+                        model=kwargs.get("model") or "",
+                        llm_provider="",
+                    )
+                messages = prompt_value
         elif (
             call_type == CallTypes.moderation.value
             or call_type == CallTypes.amoderation.value
         ):
-            messages = args[1] if len(args) > 1 else kwargs["input"]
+            if len(args) > 1:
+                messages = args[1]
+            else:
+                input_value = kwargs.get("input")
+                if input_value is None:
+                    raise BadRequestError(
+                        message="Missing required parameter: 'input' for moderation.",
+                        model=kwargs.get("model") or "",
+                        llm_provider="",
+                    )
+                messages = input_value
         elif (
             call_type == CallTypes.atext_completion.value
             or call_type == CallTypes.text_completion.value
         ):
-            messages = args[0] if len(args) > 0 else kwargs["prompt"]
+            if len(args) > 0:
+                messages = args[0]
+            else:
+                prompt_value = kwargs.get("prompt")
+                if prompt_value is None:
+                    raise BadRequestError(
+                        message="Missing required parameter: 'prompt' for text completion.",
+                        model=kwargs.get("model") or "",
+                        llm_provider="",
+                    )
+                messages = prompt_value
         elif (
             call_type == CallTypes.rerank.value or call_type == CallTypes.arerank.value
         ):
@@ -1052,7 +1082,17 @@ def function_setup(  # noqa: PLR0915
             call_type == CallTypes.atranscription.value
             or call_type == CallTypes.transcription.value
         ):
-            _file_obj: FileTypes = args[1] if len(args) > 1 else kwargs["file"]
+            if len(args) > 1:
+                _file_obj: FileTypes = args[1]
+            else:
+                file_value = kwargs.get("file")
+                if file_value is None:
+                    raise BadRequestError(
+                        message="Missing required parameter: 'file' for transcription.",
+                        model=kwargs.get("model") or "",
+                        llm_provider="",
+                    )
+                _file_obj = file_value
             # Lazy import audio_utils.utils only when needed for transcription calls
             audio_utils = _get_cached_audio_utils()
             file_checksum = audio_utils.get_audio_file_content_hash(file_obj=_file_obj)
diff --git a/tests/test_litellm/test_utils.py b/tests/test_litellm/test_utils.py
index 50dc3c6c6ec..c190e76d3b9 100644
--- a/tests/test_litellm/test_utils.py
+++ b/tests/test_litellm/test_utils.py
@@ -3865,3 +3865,60 @@ class TestValidateAndFixThinkingParam:
         validate_and_fix_thinking_param(thinking=thinking)
         assert "budgetTokens" in thinking
         assert "budget_tokens" not in thinking
+
+
+class TestFunctionSetupMissingRequiredField:
+    """
+    Regression tests: function_setup() must raise BadRequestError (status 400)
+    instead of a raw KeyError when the caller omits the required input field
+    for text-completion, image-generation, or moderation.
+    """
+
+    def _call_function_setup(self, original_function: str, **kwargs):
+        import uuid
+        from datetime import datetime
+
+        from litellm.utils import Rules, function_setup
+
+        return function_setup(
+            original_function=original_function,
+            rules_obj=Rules(),
+            start_time=datetime.now(),
+            litellm_call_id=str(uuid.uuid4()),
+            **kwargs,
+        )
+
+    def test_text_completion_missing_prompt_raises_bad_request(self):
+        with pytest.raises(litellm.BadRequestError) as exc_info:
+            self._call_function_setup(
+                original_function="atext_completion",
+                model="text-davinci-003",
+            )
+        assert exc_info.value.status_code == 400
+        assert "prompt" in str(exc_info.value)
+
+    def test_text_completion_with_prompt_succeeds(self):
+        logging_obj, kwargs = self._call_function_setup(
+            original_function="atext_completion",
+            model="text-davinci-003",
+            prompt="hello",
+        )
+        assert kwargs["prompt"] == "hello"
+
+    def test_image_generation_missing_prompt_raises_bad_request(self):
+        with pytest.raises(litellm.BadRequestError) as exc_info:
+            self._call_function_setup(
+                original_function="aimage_generation",
+                model="dall-e-3",
+            )
+        assert exc_info.value.status_code == 400
+        assert "prompt" in str(exc_info.value)
+
+    def test_moderation_missing_input_raises_bad_request(self):
+        with pytest.raises(litellm.BadRequestError) as exc_info:
+            self._call_function_setup(
+                original_function="amoderation",
+                model="text-moderation-latest",
+            )
+        assert exc_info.value.status_code == 400
+        assert "input" in str(exc_info.value)