diff --git a/litellm/llms/vertex_ai/gemini_embeddings/batch_embed_content_transformation.py b/litellm/llms/vertex_ai/gemini_embeddings/batch_embed_content_transformation.py
index 39de34169c2..834d3a5c0a8 100644
--- a/litellm/llms/vertex_ai/gemini_embeddings/batch_embed_content_transformation.py
+++ b/litellm/llms/vertex_ai/gemini_embeddings/batch_embed_content_transformation.py
@@ -312,8 +312,11 @@ def process_response(
     model_response.data = openai_embeddings
     model_response.model = model
 
-    input_text = get_formatted_prompt(data={"input": input}, call_type="embedding")
-    prompt_tokens = token_counter(model=model, text=input_text)
+    if _is_multimodal_input(input):
+        prompt_tokens = 0
+    else:
+        input_text = get_formatted_prompt(data={"input": input}, call_type="embedding")
+        prompt_tokens = token_counter(model=model, text=input_text)
     model_response.usage = Usage(
         prompt_tokens=prompt_tokens, total_tokens=prompt_tokens
     )
diff --git a/tests/test_litellm/llms/vertex_ai/gemini_embeddings/test_batch_embed_content_transformation.py b/tests/test_litellm/llms/vertex_ai/gemini_embeddings/test_batch_embed_content_transformation.py
index 2658646d59f..1417c04503c 100644
--- a/tests/test_litellm/llms/vertex_ai/gemini_embeddings/test_batch_embed_content_transformation.py
+++ b/tests/test_litellm/llms/vertex_ai/gemini_embeddings/test_batch_embed_content_transformation.py
@@ -192,4 +192,4 @@ class TestProcessResponse:
         assert len(result.data) == 2
         assert result.data[0]["index"] == 0
         assert result.data[1]["index"] == 1
-        assert result.usage.prompt_tokens >= 0
+        assert result.usage.prompt_tokens == 0