diff --git a/litellm/llms/hosted_vllm/image_edit/transformation.py b/litellm/llms/hosted_vllm/image_edit/transformation.py
index fa8c1dbfcfd..3b8cc437168 100644
--- a/litellm/llms/hosted_vllm/image_edit/transformation.py
+++ b/litellm/llms/hosted_vllm/image_edit/transformation.py
@@ -3,8 +3,17 @@ from typing import Final
 from litellm.llms.openai.image_edit.transformation import OpenAIImageEditConfig
 from litellm.secret_managers.main import get_secret_str
 
+PARAMS_VLLM_OMNI_DOES_NOT_ACCEPT: Final = frozenset({"mask", "quality", "input_fidelity"})
+
 
 class HostedVLLMImageEditConfig(OpenAIImageEditConfig):
+    def get_supported_openai_params(self, model: str) -> list:  # mutable-ok: BaseImageEditConfig contract
+        return [  # mutable-ok: BaseImageEditConfig returns list
+            param
+            for param in super().get_supported_openai_params(model)
+            if param not in PARAMS_VLLM_OMNI_DOES_NOT_ACCEPT
+        ]
+
     def validate_environment(
         self,
         headers: dict,  # mutable-ok: BaseImageEditConfig contract
diff --git a/litellm/utils.py b/litellm/utils.py
index 0f2b2f9042d..fc530f14e28 100644
--- a/litellm/utils.py
+++ b/litellm/utils.py
@@ -9239,7 +9239,7 @@ class ProviderConfigManager:
             from litellm.llms.openai.image_edit import get_openai_image_edit_config
 
             return get_openai_image_edit_config(model=model)
-        if LlmProviders.HOSTED_VLLM == provider:
+        elif LlmProviders.HOSTED_VLLM == provider:
             from litellm.llms.hosted_vllm.image_edit import get_hosted_vllm_image_edit_config
 
             return get_hosted_vllm_image_edit_config(model=model)
diff --git a/tests/test_litellm/llms/hosted_vllm/image_edit/test_hosted_vllm_image_edit_transformation.py b/tests/test_litellm/llms/hosted_vllm/image_edit/test_hosted_vllm_image_edit_transformation.py
index 5646f3d9d13..bc0ea23e249 100644
--- a/tests/test_litellm/llms/hosted_vllm/image_edit/test_hosted_vllm_image_edit_transformation.py
+++ b/tests/test_litellm/llms/hosted_vllm/image_edit/test_hosted_vllm_image_edit_transformation.py
@@ -110,3 +110,46 @@ def test_image_edit_posts_multipart_to_vllm_omni():
     assert f'name="model"\r\n\r\n{MODEL}'.encode() in request.content
     assert b'name="prompt"\r\n\r\nadd a hat' in request.content
     assert b'name="seed"\r\n\r\n42' in request.content
+
+
+@pytest.mark.parametrize("param", ["mask", "quality", "input_fidelity"])
+def test_params_vllm_omni_ignores_are_not_advertised(param: str):
+    supported = HostedVLLMImageEditConfig().get_supported_openai_params(MODEL)
+
+    assert param not in supported
+    assert {"image", "prompt", "n", "size", "response_format", "background", "user"} <= set(supported)
+
+
+def test_image_edit_rejects_quality_unless_dropped():
+    captured: list[httpx.Request] = []
+
+    def handler(request: httpx.Request) -> httpx.Response:
+        captured.append(request)
+        return httpx.Response(200, json={"created": 1712697600, "data": [{"b64_json": "aW1n"}]})
+
+    client = HTTPHandler(client=httpx.Client(transport=httpx.MockTransport(handler)))
+
+    with pytest.raises(litellm.UnsupportedParamsError, match="quality"):
+        litellm.image_edit(
+            model=f"hosted_vllm/{MODEL}",
+            image=PNG_BYTES,
+            prompt="add a hat",
+            api_base="http://localhost:8091",
+            client=client,
+            quality="low",
+        )
+    assert captured == []
+
+    litellm.image_edit(
+        model=f"hosted_vllm/{MODEL}",
+        image=PNG_BYTES,
+        prompt="add a hat",
+        api_base="http://localhost:8091",
+        client=client,
+        quality="low",
+        drop_params=True,
+    )
+
+    assert len(captured) == 1
+    assert b'name="quality"' not in captured[0].content
+    assert b'name="prompt"\r\n\r\nadd a hat' in captured[0].content