diff --git a/python/sglang/bench_serving.py b/python/sglang/bench_serving.py
index 0737fcc8c..ee981982b 100644
--- a/python/sglang/bench_serving.py
+++ b/python/sglang/bench_serving.py
@@ -493,8 +493,9 @@ def get_tokenizer(
 
 
 def get_dataset(args, tokenizer):
+    tokenize_prompt = getattr(args, "tokenize_prompt", False)
     if args.dataset_name == "sharegpt":
-        assert not args.tokenize_prompt
+        assert not tokenize_prompt
         input_requests = sample_sharegpt_requests(
             dataset_path=args.dataset_path,
             num_requests=args.num_prompts,
@@ -513,10 +514,10 @@ def get_dataset(args, tokenizer):
             tokenizer=tokenizer,
             dataset_path=args.dataset_path,
             random_sample=args.dataset_name == "random",
-            return_text=not args.tokenize_prompt,
+            return_text=not tokenize_prompt,
         )
     elif args.dataset_name == "generated-shared-prefix":
-        assert not args.tokenize_prompt
+        assert not tokenize_prompt
         input_requests = sample_generated_shared_prefix_requests(
             num_groups=args.gsp_num_groups,
             prompts_per_group=args.gsp_prompts_per_group,
@@ -527,7 +528,7 @@ def get_dataset(args, tokenizer):
             args=args,
         )
     elif args.dataset_name == "mmmu":
-        assert not args.tokenize_prompt
+        assert not tokenize_prompt
         input_requests = sample_mmmu_requests(
             num_requests=args.num_prompts,
             tokenizer=tokenizer,