diff --git a/litellm/llms/vertex_ai/ocr/deepseek_transformation.py b/litellm/llms/vertex_ai/ocr/deepseek_transformation.py
index b57a87c3325..2603552152d 100644
--- a/litellm/llms/vertex_ai/ocr/deepseek_transformation.py
+++ b/litellm/llms/vertex_ai/ocr/deepseek_transformation.py
@@ -177,9 +177,8 @@ class VertexAIDeepSeekOCRConfig(BaseOCRConfig):
             content_item = {"type": "image_url", "image_url": document_url}
 
         # Build DeepSeek OCR request
-        provider_model: Final = model if model.startswith("deepseek-ai/") else f"deepseek-ai/{model}"
         data: Final = {
-            "model": provider_model,
+            "model": "deepseek-ai/" + model,
             "messages": [{"role": "user", "content": [content_item]}],
         }
 
diff --git a/litellm/ocr/main.py b/litellm/ocr/main.py
index f3ecb545f66..b918f013700 100644
--- a/litellm/ocr/main.py
+++ b/litellm/ocr/main.py
@@ -191,10 +191,6 @@ def _prepare_ocr_request(
 def _rust_ocr_supported(prepared_request: _PreparedOCRRequest) -> bool:
     if prepared_request.optional_params.get(OCR_REQUEST_FORMAT_PARAM) == "native":
         return False
-    if prepared_request.extra_headers is not None and any(
-        not isinstance(value, str) for value in prepared_request.extra_headers.values()
-    ):
-        return False
     return prepared_request.custom_llm_provider in _RUST_OCR_PROVIDERS
 
 
diff --git a/tests/test_litellm/ocr/test_fixture_models.py b/tests/test_litellm/ocr/test_fixture_models.py
index 34856a7bee1..24fb946e9bb 100644
--- a/tests/test_litellm/ocr/test_fixture_models.py
+++ b/tests/test_litellm/ocr/test_fixture_models.py
@@ -457,19 +457,6 @@ def test_unqualified_reducto_models_accept_explicit_provider(
     assert sdk_input.model == model
 
 
-@pytest.mark.parametrize("model", ("deepseek-ocr-maas", "deepseek-ai/deepseek-ocr-maas"))
-def test_vertex_deepseek_request_uses_single_provider_namespace(model: str) -> None:
-    request: Final = VertexAIDeepSeekOCRConfig().transform_ocr_request(  # pyright: ignore[reportUnknownMemberType]
-        model=model,
-        document={"type": "image_url", "image_url": "data:image/png;base64,AA=="},
-        optional_params={},
-        headers={},
-    )
-
-    data: Final = cast(dict[str, object], request.data)
-    assert data["model"] == "deepseek-ai/deepseek-ocr-maas"
-
-
 @pytest.mark.parametrize(
     "document",
     (
diff --git a/tests/test_litellm/ocr/test_sdk_parity.py b/tests/test_litellm/ocr/test_sdk_parity.py
index edce2840a0a..a60e536a3ac 100644
--- a/tests/test_litellm/ocr/test_sdk_parity.py
+++ b/tests/test_litellm/ocr/test_sdk_parity.py
@@ -147,15 +147,6 @@ INVALID_OCR_CASES: Final = (
         expected_message="Invalid `features` for Azure Document Intelligence",
         extra_kwargs=(("features", [1]),),
     ),
-    InvalidOcrCase(
-        name="invalid_header_value",
-        model="mistral/mistral-ocr-latest",
-        document={"type": "document_url", "document_url": "data:application/pdf;base64,AA=="},
-        expected_exception_type="litellm.exceptions.InternalServerError",
-        expected_status_code=500,
-        expected_message="Header value must be str or bytes",
-        extra_kwargs=(("extra_headers", {"x-invalid": 1}),),
-    ),
 )