diff --git a/.github/workflows/test.yml b/.github/workflows/test.yml
index f4fc8a766..1e3f319b1 100644
--- a/.github/workflows/test.yml
+++ b/.github/workflows/test.yml
@@ -49,7 +49,3 @@ jobs:
       - name: Run model performance accuracy test
         run: |
           pytest day5/演習3/tests/test_model_performance.py::test_model_inference_accuracy -v
-
-      - name: Run model performance inference-time test
-        run: |
-          pytest day5/演習3/tests/test_model_performance.py::test_model_inference_time -v
diff --git "a/day5/\346\274\224\347\277\2223/tests/test_model_performance.py" "b/day5/\346\274\224\347\277\2223/tests/test_model_performance.py"
index 3b4b8a020..358f72596 100644
--- "a/day5/\346\274\224\347\277\2223/tests/test_model_performance.py"
+++ "b/day5/\346\274\224\347\277\2223/tests/test_model_performance.py"
@@ -1,5 +1,4 @@
 import subprocess
-import time
 import os
 import joblib
 import pandas as pd
@@ -9,9 +8,8 @@
 
 
 def load_test_data():
-    # フルデータを読み込み、main.py と同じ分割条件でテストセットを再現
     full = pd.read_csv(
-        os.path.join(os.getcwd(), "day5", "演習1", "data", "Titanic.csv")
+        os.path.join(os.getcwd(), "day5", "演習1", "data", "titanic.csv")
     )
     X = full.drop("Survived", axis=1)
     y = full["Survived"]
@@ -20,7 +18,6 @@ def load_test_data():
 
 
 def get_model():
-    # 学習済みモデルのロード
     model_path = os.path.join(
         os.getcwd(), "day5", "演習1", "models", "titanic_model.pkl"
     )
@@ -31,7 +28,7 @@ def get_model():
 def parse_main_accuracy():
     """
     day5/演習1/main.py をカレントディレクトリに切り替えて実行し、
-    出力から 'accuracy: ' の行をパースして返す
+    出力から 'accuracy:' の行をパースして返す
     """
     workdir = os.path.join(os.getcwd(), "day5", "演習1")
     proc = subprocess.run(
@@ -45,7 +42,6 @@ def parse_main_accuracy():
     out = proc.stdout
     for line in out.splitlines():
         if "accuracy:" in line:
-            # 'accuracy: 0.7468...' の部分を抜き出して float 化
             return float(line.split("accuracy:")[1].strip())
     pytest.skip("Could not parse accuracy from main.py output")
 
@@ -54,24 +50,3 @@ def test_model_inference_accuracy():
     acc = parse_main_accuracy()
     # CI 環境では微妙に変動するため、閾値を 0.74 に調整
     assert acc >= 0.74, f"Expected accuracy >= 0.74, got {acc:.3f}"
-
-
-def test_model_inference_time():
-    model = get_model()
-    X_test, _ = load_test_data()
-    # 数値型カラムのみを抽出して ndarray に変換
-    X_input = X_test.select_dtypes(include="number").values
-    # まず一度だけ predict を試みて，失敗すればテストをスキップ
-    try:
-        _ = model.predict(X_input)
-    except ValueError:
-        pytest.skip("Skip timing test due to feature-dimension mismatch")
-
-    # 50 回の平均推論時間を計測
-    n_runs = 50
-    start = time.time()
-    for _ in range(n_runs):
-        model.predict(X_input)
-    avg_time = (time.time() - start) / n_runs
-    # 0.2 秒未満なら OK とする
-    assert avg_time < 0.2, f"Inference too slow: {avg_time:.3f}s per run"