From a1cac484778dfddb5dbc281c0a0e66797b682559 Mon Sep 17 00:00:00 2001
From: Woosuk Kwon <woosuk@thinkingmachines.ai>
Date: Thu, 23 Oct 2025 05:40:02 +0000
Subject: [PATCH] Turn off usage

---
 vllm/benchmarks/lib/endpoint_request_func.py | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/vllm/benchmarks/lib/endpoint_request_func.py b/vllm/benchmarks/lib/endpoint_request_func.py
index 4f427a31b9ee1..932984fb1d05b 100644
--- a/vllm/benchmarks/lib/endpoint_request_func.py
+++ b/vllm/benchmarks/lib/endpoint_request_func.py
@@ -165,9 +165,10 @@ async def async_request_openai_completions(
         "max_tokens": request_func_input.output_len,
         "logprobs": request_func_input.logprobs,
         "stream": True,
-        "stream_options": {
-            "include_usage": True,
-        },
+        # NOTE(woosuk): Disabled for PD.
+        # "stream_options": {
+        #     "include_usage": True,
+        # },
     }
     _update_payload_common(payload, request_func_input)