diff --git a/litellm/llms/openai/chat/gpt_5_transformation.py b/litellm/llms/openai/chat/gpt_5_transformation.py
index d0e5ff01e71..c8504b54b19 100644
--- a/litellm/llms/openai/chat/gpt_5_transformation.py
+++ b/litellm/llms/openai/chat/gpt_5_transformation.py
@@ -311,6 +311,20 @@ class OpenAIGPT5Config(OpenAIGPTConfig):
                         message=(f"reasoning_effort={effective_effort} is not supported for this model."),
                         status_code=400,
                     )
+        elif effective_effort == "none":
+            # none is opt-out: unknown models pass through; only block when
+            # the model map explicitly sets supports_none_reasoning_effort=false.
+            # gpt-5 and gpt-5-mini are both marked supports_none_reasoning_effort=false
+            # and OpenAI returns 400 for none on those models.
+            if self._is_reasoning_effort_level_explicitly_disabled(model, effective_effort):
+                if litellm.drop_params or drop_params:
+                    non_default_params.pop("reasoning_effort", None)
+                    optional_params.pop("reasoning_effort", None)
+                else:
+                    raise litellm.utils.UnsupportedParamsError(
+                        message=(f"reasoning_effort={effective_effort} is not supported for this model."),
+                        status_code=400,
+                    )
 
         ################################################################
         # max_tokens is not supported for gpt-5 models on OpenAI API