diff --git a/benchmark/benchmark.py b/benchmark/benchmark.py
index dbf62aacc..9d1717a52 100755
--- a/benchmark/benchmark.py
+++ b/benchmark/benchmark.py
@@ -820,7 +820,9 @@ def cleanup_test_output(output, testdir):
     # remove timing info, to avoid randomizing the response to GPT
     res = re.sub(
         r"\bin \d+\.\d+s\b",
-
+        "",
+        output
+    )
     res = res.replace(str(testdir), str(testdir.name))
     return res