diff --git a/sagemaker-train/tests/integ/train/test_dpo_trainer_integration.py b/sagemaker-train/tests/integ/train/test_dpo_trainer_integration.py
index 69b99e7584..d220e77aa9 100644
--- a/sagemaker-train/tests/integ/train/test_dpo_trainer_integration.py
+++ b/sagemaker-train/tests/integ/train/test_dpo_trainer_integration.py
@@ -19,8 +19,10 @@
 from sagemaker.core.helper.session_helper import Session
 from sagemaker.train.dpo_trainer import DPOTrainer
 from sagemaker.train.common import TrainingType
+import pytest
 
 
+@pytest.mark.skip(reason="Skipping GPU resource intensive test")
 def test_dpo_trainer_lora_complete_workflow(sagemaker_session):
     """Test complete DPO training workflow with LORA."""
     # Create DPOTrainer instance with comprehensive configuration
@@ -61,6 +63,7 @@ def test_dpo_trainer_lora_complete_workflow(sagemaker_session):
     assert training_job.output_model_package_arn is not None
 
 
+@pytest.mark.skip(reason="Skipping GPU resource intensive test")
 def test_dpo_trainer_with_validation_dataset(sagemaker_session):
     """Test DPO trainer with both training and validation datasets."""
     
diff --git a/sagemaker-train/tests/integ/train/test_rlaif_trainer_integration.py b/sagemaker-train/tests/integ/train/test_rlaif_trainer_integration.py
index bdb6168332..9f3594ad01 100644
--- a/sagemaker-train/tests/integ/train/test_rlaif_trainer_integration.py
+++ b/sagemaker-train/tests/integ/train/test_rlaif_trainer_integration.py
@@ -18,8 +18,10 @@
 from sagemaker.core.helper.session_helper import Session
 from sagemaker.train.rlaif_trainer import RLAIFTrainer
 from sagemaker.train.common import TrainingType
+import pytest
 
 
+@pytest.mark.skip(reason="Skipping GPU resource intensive test")
 def test_rlaif_trainer_lora_complete_workflow(sagemaker_session):
     """Test complete RLAIF training workflow with LORA."""
     
@@ -59,6 +61,7 @@ def test_rlaif_trainer_lora_complete_workflow(sagemaker_session):
     assert training_job.output_model_package_arn is not None
 
 
+@pytest.mark.skip(reason="Skipping GPU resource intensive test")
 def test_rlaif_trainer_with_custom_reward_settings(sagemaker_session):
     """Test RLAIF trainer with different reward model and prompt."""
     
@@ -97,6 +100,7 @@ def test_rlaif_trainer_with_custom_reward_settings(sagemaker_session):
     assert training_job.output_model_package_arn is not None
 
 
+@pytest.mark.skip(reason="Skipping GPU resource intensive test")
 def test_rlaif_trainer_continued_finetuning(sagemaker_session):
     """Test complete RLAIF training workflow with LORA."""
 
diff --git a/sagemaker-train/tests/integ/train/test_rlvr_trainer_integration.py b/sagemaker-train/tests/integ/train/test_rlvr_trainer_integration.py
index bc13181d50..d723b3338c 100644
--- a/sagemaker-train/tests/integ/train/test_rlvr_trainer_integration.py
+++ b/sagemaker-train/tests/integ/train/test_rlvr_trainer_integration.py
@@ -22,6 +22,7 @@
 from sagemaker.train.common import TrainingType
 
 
+@pytest.mark.skip(reason="Skipping GPU resource intensive test")
 def test_rlvr_trainer_lora_complete_workflow(sagemaker_session):
     """Test complete RLVR training workflow with LORA."""
     
@@ -59,6 +60,7 @@ def test_rlvr_trainer_lora_complete_workflow(sagemaker_session):
     assert training_job.output_model_package_arn is not None
 
 
+@pytest.mark.skip(reason="Skipping GPU resource intensive test")
 def test_rlvr_trainer_with_custom_reward_function(sagemaker_session):
     """Test RLVR trainer with custom reward function."""
     
@@ -96,7 +98,8 @@ def test_rlvr_trainer_with_custom_reward_function(sagemaker_session):
     assert training_job.output_model_package_arn is not None
 
 
-@pytest.mark.skipif(os.environ.get('AWS_DEFAULT_REGION') != 'us-east-1', reason="Nova models only available in us-east-1")
+# @pytest.mark.skipif(os.environ.get('AWS_DEFAULT_REGION') != 'us-east-1', reason="Nova models only available in us-east-1")
+@pytest.mark.skip(reason="Skipping GPU resource intensive test")
 def test_rlvr_trainer_nova_workflow(sagemaker_session):
     """Test RLVR training workflow with Nova model."""
     import os
diff --git a/sagemaker-train/tests/integ/train/test_sft_trainer_integration.py b/sagemaker-train/tests/integ/train/test_sft_trainer_integration.py
index c5dd66f7b8..e473761bed 100644
--- a/sagemaker-train/tests/integ/train/test_sft_trainer_integration.py
+++ b/sagemaker-train/tests/integ/train/test_sft_trainer_integration.py
@@ -22,6 +22,7 @@
 from sagemaker.train.common import TrainingType
 
 
+@pytest.mark.skip(reason="Skipping GPU resource intensive test")
 def test_sft_trainer_lora_complete_workflow(sagemaker_session):
     """Test complete SFT training workflow with LORA."""
     
@@ -57,6 +58,7 @@ def test_sft_trainer_lora_complete_workflow(sagemaker_session):
     assert training_job.output_model_package_arn is not None
 
 
+@pytest.mark.skip(reason="Skipping GPU resource intensive test")
 def test_sft_trainer_with_validation_dataset(sagemaker_session):
     """Test SFT trainer with both training and validation datasets."""
 
@@ -90,7 +92,8 @@ def test_sft_trainer_with_validation_dataset(sagemaker_session):
     assert hasattr(training_job, 'output_model_package_arn')
 
 
-@pytest.mark.skipif(os.environ.get('AWS_DEFAULT_REGION') != 'us-east-1', reason="Nova models only available in us-east-1")
+# @pytest.mark.skipif(os.environ.get('AWS_DEFAULT_REGION') != 'us-east-1', reason="Nova models only available in us-east-1")
+@pytest.mark.skip(reason="Skipping GPU resource intensive test")
 def test_sft_trainer_nova_workflow(sagemaker_session):
     """Test SFT trainer with Nova model."""
     import os