diff --git a/llama_stack/providers/remote/inference/nvidia/nvidia.py b/llama_stack/providers/remote/inference/nvidia/nvidia.py
index 7052cfb57..ec4cba742 100644
--- a/llama_stack/providers/remote/inference/nvidia/nvidia.py
+++ b/llama_stack/providers/remote/inference/nvidia/nvidia.py
@@ -7,7 +7,7 @@
 import warnings
 from collections.abc import AsyncIterator
 
-from openai import NOT_GIVEN, APIConnectionError, BadRequestError
+from openai import NOT_GIVEN, APIConnectionError
 
 from llama_stack.apis.common.content_types import (
     InterleavedContent,
@@ -197,15 +197,11 @@ class NVIDIAInferenceAdapter(OpenAIMixin, Inference, ModelRegistryHelper):
             }
             extra_body["input_type"] = task_type_options[task_type]
 
-        try:
-            response = await self.client.embeddings.create(
-                model=provider_model_id,
-                input=input,
-                extra_body=extra_body,
-            )
-        except BadRequestError as e:
-            raise ValueError(f"Failed to get embeddings: {e}") from e
-
+        response = await self.client.embeddings.create(
+            model=provider_model_id,
+            input=input,
+            extra_body=extra_body,
+        )
         #
         # OpenAI: CreateEmbeddingResponse(data=[Embedding(embedding=list[float], ...)], ...)
         #  ->
diff --git a/tests/integration/inference/test_embedding.py b/tests/integration/inference/test_embedding.py
index 075f927f7..e592a6b14 100644
--- a/tests/integration/inference/test_embedding.py
+++ b/tests/integration/inference/test_embedding.py
@@ -55,7 +55,7 @@
 #
 
 import pytest
-from llama_stack_client import BadRequestError
+from llama_stack_client import BadRequestError as LlamaStackBadRequestError
 from llama_stack_client.types import EmbeddingsResponse
 from llama_stack_client.types.shared.interleaved_content import (
     ImageContentItem,
@@ -63,6 +63,9 @@ from llama_stack_client.types.shared.interleaved_content import (
     ImageContentItemImageURL,
     TextContentItem,
 )
+from openai import BadRequestError as OpenAIBadRequestError
+
+from llama_stack.core.library_client import LlamaStackAsLibraryClient
 
 DUMMY_STRING = "hello"
 DUMMY_STRING2 = "world"
@@ -203,7 +206,14 @@ def test_embedding_truncation_error(
 ):
     if inference_provider_type not in SUPPORTED_PROVIDERS:
         pytest.xfail(f"{inference_provider_type} doesn't support embedding model yet")
-    with pytest.raises(BadRequestError):
+    # Using LlamaStackClient from llama_stack_client will raise llama_stack_client.BadRequestError
+    # While using LlamaStackAsLibraryClient from llama_stack.distribution.library_client will raise the error that the backend raises
+    error_type = (
+        OpenAIBadRequestError
+        if isinstance(llama_stack_client, LlamaStackAsLibraryClient)
+        else LlamaStackBadRequestError
+    )
+    with pytest.raises(error_type):
         llama_stack_client.inference.embeddings(
             model_id=embedding_model_id,
             contents=[DUMMY_LONG_TEXT],
@@ -283,7 +293,8 @@ def test_embedding_text_truncation_error(
 ):
     if inference_provider_type not in SUPPORTED_PROVIDERS:
         pytest.xfail(f"{inference_provider_type} doesn't support embedding model yet")
-    with pytest.raises(BadRequestError):
+    error_type = ValueError if isinstance(llama_stack_client, LlamaStackAsLibraryClient) else LlamaStackBadRequestError
+    with pytest.raises(error_type):
         llama_stack_client.inference.embeddings(
             model_id=embedding_model_id,
             contents=[DUMMY_STRING],