diff --git a/litellm/llms/groq/chat/transformation.py b/litellm/llms/groq/chat/transformation.py
index 165301efb5c..20e0d412edc 100644
--- a/litellm/llms/groq/chat/transformation.py
+++ b/litellm/llms/groq/chat/transformation.py
@@ -1,9 +1,27 @@
 """
 Translate from OpenAI's `/v1/chat/completions` to Groq's `/v1/chat/completions`
 """
-from typing import Any, Coroutine, List, Literal, Optional, Tuple, Union, cast, overload
+from typing import (
+    Any,
+    Coroutine,
+    List,
+    Literal,
+    Optional,
+    Tuple,
+    Union,
+    cast,
+    overload,
+    Iterator,
+    AsyncIterator,
+)
 
 import httpx
+
+from litellm.llms.openai.chat.gpt_transformation import (
+    OpenAIChatCompletionStreamingHandler,
+)
+from litellm.llms.openai.common_utils import OpenAIError
+
 from pydantic import BaseModel
 
 import litellm
@@ -16,7 +34,7 @@ from litellm.types.llms.openai import (
     ChatCompletionToolParam,
     ChatCompletionToolParamFunctionChunk,
 )
-from litellm.types.utils import ModelResponse
+from litellm.types.utils import ModelResponse, ModelResponseStream
 
 from ...openai_like.chat.transformation import OpenAILikeChatConfig
 
@@ -65,6 +83,18 @@ class GroqChatConfig(OpenAILikeChatConfig):
     def get_config(cls):
         return super().get_config()
 
+    def get_model_response_iterator(
+        self,
+        streaming_response: Union[Iterator[str], AsyncIterator[str], ModelResponse],
+        sync_stream: bool,
+        json_mode: Optional[bool] = False,
+    ) -> Any:
+        return GroqChatCompletionStreamingHandler(
+            streaming_response=streaming_response,
+            sync_stream=sync_stream,
+            json_mode=json_mode,
+        )
+
     def get_supported_openai_params(self, model: str) -> list:
         base_params = super().get_supported_openai_params(model)
         try:
@@ -209,7 +239,6 @@ class GroqChatConfig(OpenAILikeChatConfig):
         )
 
         return optional_params
-    
 
     def transform_response(
         self,
@@ -239,12 +268,17 @@ class GroqChatConfig(OpenAILikeChatConfig):
             json_mode=json_mode,
         )
 
-        mapped_service_tier: Literal["auto", "default", "flex"] = self._map_groq_service_tier(original_service_tier=getattr(model_response, "service_tier"))
+        mapped_service_tier: Literal[
+            "auto", "default", "flex"
+        ] = self._map_groq_service_tier(
+            original_service_tier=getattr(model_response, "service_tier")
+        )
         setattr(model_response, "service_tier", mapped_service_tier)
         return model_response
-    
 
-    def _map_groq_service_tier(self, original_service_tier: Optional[str]) -> Literal["auto", "default", "flex"]:
+    def _map_groq_service_tier(
+        self, original_service_tier: Optional[str]
+    ) -> Literal["auto", "default", "flex"]:
         """
         Ensure groq service tier is OpenAI compatible.
         """
@@ -252,5 +286,16 @@ class GroqChatConfig(OpenAILikeChatConfig):
             return "auto"
         if original_service_tier not in ["auto", "default", "flex"]:
             return "auto"
-        
-        return cast(Literal["auto", "default", "flex"], original_service_tier)
\ No newline at end of file
+
+        return cast(Literal["auto", "default", "flex"], original_service_tier)
+
+
+class GroqChatCompletionStreamingHandler(OpenAIChatCompletionStreamingHandler):
+    def chunk_parser(self, chunk: dict) -> ModelResponseStream:
+        error = chunk.get("error")
+        if error:
+            raise OpenAIError(
+                status_code=error.get("code"), message=error.get("message"), body=error
+            )
+
+        return super().chunk_parser(chunk)