diff --git a/qa/L0_dlpack_multi_gpu/test.sh b/qa/L0_dlpack_multi_gpu/test.sh
index 65dd2ad0b3..0aaf135b5f 100755
--- a/qa/L0_dlpack_multi_gpu/test.sh
+++ b/qa/L0_dlpack_multi_gpu/test.sh
@@ -1,5 +1,5 @@
 #!/bin/bash
-# Copyright 2023-2025, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# Copyright 2023-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
 #
 # Redistribution and use in source and binary forms, with or without
 # modification, are permitted provided that the following conditions
@@ -41,7 +41,7 @@ source ../common/util.sh
 
 # Uninstall the non CUDA version of PyTorch
 pip3 uninstall -y torch
-pip3 install torch==2.3.1+cu118 -f https://download.pytorch.org/whl/torch_stable.html
+pip3 install torch -f https://download.pytorch.org/whl/cu130
 
 # Install CuPy for testing non_blocking compute streams
 pip3 install cupy-cuda13x
diff --git a/qa/L0_io/test.sh b/qa/L0_io/test.sh
index 10318da4db..43c7e7dd41 100755
--- a/qa/L0_io/test.sh
+++ b/qa/L0_io/test.sh
@@ -1,5 +1,5 @@
 #!/bin/bash
-# Copyright 2019-2025, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# Copyright 2019-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
 #
 # Redistribution and use in source and binary forms, with or without
 # modification, are permitted provided that the following conditions
@@ -55,7 +55,7 @@ LD_LIBRARY_PATH=/opt/tritonserver/lib:$LD_LIBRARY_PATH
 rm -f $CLIENT_LOG*
 
 # PyTorch is required for the Python backend dlpack add sub models
-pip3 install torch==2.3.1+cu118 -f https://download.pytorch.org/whl/torch_stable.html
+pip3 install torch -f https://download.pytorch.org/whl/cu130
 RET=0
 
 # Prepare float32 models with basic config
diff --git a/qa/L0_libtorch_instance_group_kind_model/test.sh b/qa/L0_libtorch_instance_group_kind_model/test.sh
index 9cbb6bf550..a68a8b3257 100755
--- a/qa/L0_libtorch_instance_group_kind_model/test.sh
+++ b/qa/L0_libtorch_instance_group_kind_model/test.sh
@@ -1,5 +1,5 @@
 #!/bin/bash
-# Copyright 2023-2024, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# Copyright 2023-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
 #
 # Redistribution and use in source and binary forms, with or without
 # modification, are permitted provided that the following conditions
@@ -39,7 +39,7 @@ if [ ! -z "$TEST_REPO_ARCH" ]; then
 fi
 
 pip3 uninstall -y torch
-pip3 install torch==2.3.1+cu118 -f https://download.pytorch.org/whl/torch_stable.html
+pip3 install torch -f https://download.pytorch.org/whl/cu130
 
 DATADIR=/data/inferenceserver/${REPO_VERSION}/qa_model_repository
 SERVER=/opt/tritonserver/bin/tritonserver
diff --git a/qa/L0_warmup/test.sh b/qa/L0_warmup/test.sh
index c3e6885062..84a0752229 100755
--- a/qa/L0_warmup/test.sh
+++ b/qa/L0_warmup/test.sh
@@ -1,5 +1,5 @@
 #!/bin/bash
-# Copyright 2019-2025, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# Copyright 2019-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
 #
 # Redistribution and use in source and binary forms, with or without
 # modification, are permitted provided that the following conditions
@@ -420,7 +420,7 @@ wait $SERVER_PID
 # Test the onnx model to verify that the memory type of the output tensor
 # remains unchanged with the warmup setting
 pip3 uninstall -y torch
-pip3 install torch==2.3.1+cu118 -f https://download.pytorch.org/whl/torch_stable.html
+pip3 install torch -f https://download.pytorch.org/whl/cu130
 
 rm -fr models && mkdir models
 cp -r /data/inferenceserver/${REPO_VERSION}/qa_model_repository/onnx_nobatch_float32_float32_float32 models/.