diff --git a/test/srt/run_suite.py b/test/srt/run_suite.py
index 1c4766e24..0f4adffad 100644
--- a/test/srt/run_suite.py
+++ b/test/srt/run_suite.py
@@ -103,6 +103,7 @@ suites = {
         TestFile("test_original_logprobs.py", 41),
         TestFile("test_page_size.py", 60),
         TestFile("test_penalty.py", 82),
+        TestFile("test_piecewise_cuda_graph.py", 180),
         TestFile("test_priority_scheduling.py", 130),
         TestFile("test_pytorch_sampling_backend.py", 66),
         TestFile("test_radix_attention.py", 105),
diff --git a/test/srt/test_piecewise_cuda_graph.py b/test/srt/test_piecewise_cuda_graph.py
index 4b0042dbc..d50c67453 100644
--- a/test/srt/test_piecewise_cuda_graph.py
+++ b/test/srt/test_piecewise_cuda_graph.py
@@ -1,8 +1,5 @@
-import time
 import unittest
 
-import requests
-
 from sglang.srt.utils import kill_process_tree
 from sglang.test.run_eval import run_eval
 from sglang.test.test_utils import (
@@ -32,18 +29,6 @@ class TestPiecewiseCudaGraphCorrectness(CustomTestCase):
     def tearDownClass(cls):
         kill_process_tree(cls.process.pid)
 
-    def test_gpqa(self):
-        args = SimpleNamespace(
-            base_url=self.base_url,
-            model=self.model,
-            eval_name="gpqa",
-            num_examples=64,
-            num_threads=16,
-        )
-
-        metrics = run_eval(args)
-        self.assertGreaterEqual(metrics["score"], 0.235)
-
     def test_mmlu(self):
         args = SimpleNamespace(
             base_url=self.base_url,