diff --git a/recipe/bld.bat b/recipe/bld.bat
index 30cc5d4f..b785aa32 100644
--- a/recipe/bld.bat
+++ b/recipe/bld.bat
@@ -16,7 +16,8 @@ if "%build_with_cuda%" == "" goto cuda_flags_end
 
 set CUDA_PATH=C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v%desired_cuda%
 set CUDA_BIN_PATH=%CUDA_PATH%\bin
-set TORCH_CUDA_ARCH_LIST=5.0;6.0;6.1;7.0;7.5;8.0;8.6;8.9;9.0+PTX
+REM Keep this list in sync with https://github.com/pytorch/pytorch/blob/07fa6e2c8b003319f85a469307f1b1dd73f6026c/.ci/manywheel/build_cuda.sh#L60
+set TORCH_CUDA_ARCH_LIST=5.0;6.0;7.0;7.5;8.0;8.6;9.0+PTX
 set TORCH_NVCC_FLAGS=-Xfatbin -compress-all
 
 :cuda_flags_end
diff --git a/recipe/build.sh b/recipe/build.sh
index 19da0ad4..851f5c58 100644
--- a/recipe/build.sh
+++ b/recipe/build.sh
@@ -168,7 +168,8 @@ elif [[ ${cuda_compiler_version} != "None" ]]; then
     esac
     case ${cuda_compiler_version} in
         12.6)
-            export TORCH_CUDA_ARCH_LIST="5.0;6.0;6.1;7.0;7.5;8.0;8.6;8.9;9.0+PTX"
+            # Keep this list in sync with https://github.com/pytorch/pytorch/blob/07fa6e2c8b003319f85a469307f1b1dd73f6026c/.ci/manywheel/build_cuda.sh#L60
+            export TORCH_CUDA_ARCH_LIST="5.0;6.0;7.0;7.5;8.0;8.6;9.0+PTX"
             ;;
         *)
             echo "unsupported cuda version. edit build.sh"