diff --git a/.azure-pipelines/scripts/install_nc.sh b/.azure-pipelines/scripts/install_nc.sh
index 8f6f7b4800b..e16ab09b87e 100644
--- a/.azure-pipelines/scripts/install_nc.sh
+++ b/.azure-pipelines/scripts/install_nc.sh
@@ -9,8 +9,8 @@ if [[ $1 = *"3x_pt"* ]]; then
         python setup.py pt bdist_wheel
     else
         echo -e "\n Install torch CPU ... "
-        pip install torch==2.8.0 torchvision --index-url https://download.pytorch.org/whl/cpu
-        python -m pip install intel-extension-for-pytorch==2.8.0 oneccl_bind_pt --index-url https://pytorch-extension.intel.com/release-whl/stable/cpu/us/
+        pip install torch==2.9.0 torchvision --index-url https://download.pytorch.org/whl/cpu
+        # python -m pip install intel-extension-for-pytorch==2.8.0 oneccl_bind_pt --index-url https://pytorch-extension.intel.com/release-whl/stable/cpu/us/
         python -m pip install --no-cache-dir -r requirements.txt
         python setup.py bdist_wheel
     fi
diff --git a/test/3x/torch/algorithms/smooth_quant/test_sq_utility.py b/test/3x/torch/algorithms/smooth_quant/test_sq_utility.py
index 0a645e27772..6748d3dda39 100644
--- a/test/3x/torch/algorithms/smooth_quant/test_sq_utility.py
+++ b/test/3x/torch/algorithms/smooth_quant/test_sq_utility.py
@@ -2,6 +2,7 @@
 
 import pytest
 import torch
+from packaging.version import Version
 
 
 class Model(torch.nn.Module):
@@ -20,7 +21,7 @@ def forward(self, x):
 
 model = Model()
 
-
+@pytest.mark.skipif(not Version(torch.__version__) < Version("2.9.0"), reason="only for torch<2.9.0 [ipex]")
 def test_captured_dataloader():
     from neural_compressor.torch.algorithms.smooth_quant import build_captured_dataloader
 
diff --git a/test/3x/torch/quantization/test_autoround.py b/test/3x/torch/quantization/test_autoround.py
index 1a40c0442ab..3bda8cb7eae 100644
--- a/test/3x/torch/quantization/test_autoround.py
+++ b/test/3x/torch/quantization/test_autoround.py
@@ -173,7 +173,7 @@ def test_conv1d(self):
         tokenizer =  AutoTokenizer.from_pretrained("MBZUAI/LaMini-GPT-124M", trust_remote_code=True)
         text = "Replace me by any text you'd like."
         encoded_input = tokenizer(text, return_tensors="pt")
-        quant_config = AutoRoundConfig(nsamples=32, seqlen=10, iters=0,tokenizer=tokenizer,export_format="auto_round")
+        quant_config = AutoRoundConfig(nsamples=32, seqlen=10, iters=0, amp=False, tokenizer=tokenizer, export_format="auto_round")
         model = prepare(model=model, quant_config=quant_config)
         q_model = convert(model)
         output = tokenizer.decode(q_model.generate(**encoded_input, max_new_tokens=10)[0])
diff --git a/test/3x/torch/quantization/test_smooth_quant.py b/test/3x/torch/quantization/test_smooth_quant.py
index 147c1adfe1d..ea8c15ddfbe 100644
--- a/test/3x/torch/quantization/test_smooth_quant.py
+++ b/test/3x/torch/quantization/test_smooth_quant.py
@@ -4,6 +4,7 @@
 
 import pytest
 import torch
+from packaging.version import Version
 
 from neural_compressor.torch.quantization import SmoothQuantConfig, convert, get_default_sq_config, prepare, quantize
 from neural_compressor.torch.utils import is_ipex_available
@@ -33,8 +34,7 @@ def forward(self, x):
 def run_fn(model):
     for i in range(10):
         model(example_inputs)
-
-
+@pytest.mark.skipif(not Version(torch.__version__) < Version("2.9.0"), reason="only for torch<2.9.0 [ipex]")
 class TestSmoothQuant:
     def teardown_class(self):
         shutil.rmtree("saved_results", ignore_errors=True)
diff --git a/test/3x/torch/quantization/test_static_quant.py b/test/3x/torch/quantization/test_static_quant.py
index a03b2939425..29f506040af 100644
--- a/test/3x/torch/quantization/test_static_quant.py
+++ b/test/3x/torch/quantization/test_static_quant.py
@@ -4,6 +4,7 @@
 
 import pytest
 import torch
+from packaging.version import Version
 
 try:
     import intel_extension_for_pytorch as ipex
@@ -52,7 +53,7 @@ def run_fn(model):
     model(torch.rand((1, 30)))
     model(torch.rand((1, 30)))
 
-
+@pytest.mark.skipif(not Version(torch.__version__) < Version("2.9.0"), reason="only for torch<2.9.0 [ipex]")
 class TestStaticQuant:
     def setup_class(self):
         self.fp32_model = build_simple_torch_model()
diff --git a/test/3x/torch/quantization/weight_only/test_transformers.py b/test/3x/torch/quantization/weight_only/test_transformers.py
index 6310fdafcef..6c291f81e35 100644
--- a/test/3x/torch/quantization/weight_only/test_transformers.py
+++ b/test/3x/torch/quantization/weight_only/test_transformers.py
@@ -23,6 +23,7 @@
 
 ipex_version = get_ipex_version()
 
+
 try:
     import auto_round
 
@@ -30,6 +31,7 @@
 except ImportError:
     auto_round_installed = False
 
+@pytest.mark.skipif(not Version(torch.__version__) < Version("2.9.0"), reason="only for torch<2.9.0 [ipex]")
 class TestTansformersLikeAPI:
     def setup_class(self):
         self.model_name_or_path = "hf-tiny-model-private/tiny-random-GPTJForCausalLM"