huggingface · danieldk · Sep 29, 2026 · Sep 28, 2026
diff --git a/Cargo.lock b/Cargo.lock
diff --git a/docs/source/builder/build-variants.md b/docs/source/builder/build-variants.md
@@ -17,43 +17,43 @@ available. This list will be updated as new PyTorch versions are released.
 
 ## CPU aarch64-linux
 
-- `torch213-cxx11-cpu-aarch64-linux`
-- `torch214-cxx11-cpu-aarch64-linux`
+- `torch213-cpu-aarch64-linux`
+- `torch214-cpu-aarch64-linux`
 
 ## CUDA aarch64-linux
 
-- `torch213-cxx11-cu126-aarch64-linux`
-- `torch213-cxx11-cu130-aarch64-linux`
-- `torch213-cxx11-cu132-aarch64-linux`
-- `torch214-cxx11-cu126-aarch64-linux`
-- `torch214-cxx11-cu130-aarch64-linux`
-- `torch214-cxx11-cu132-aarch64-linux`
+- `torch213-cu126-aarch64-linux`
+- `torch213-cu130-aarch64-linux`
+- `torch213-cu132-aarch64-linux`
+- `torch214-cu126-aarch64-linux`
+- `torch214-cu130-aarch64-linux`
+- `torch214-cu132-aarch64-linux`
 
 ## CPU x86_64-linux
 
-- `torch213-cxx11-cpu-x86_64-linux`
-- `torch214-cxx11-cpu-x86_64-linux`
+- `torch213-cpu-x86_64-linux`
+- `torch214-cpu-x86_64-linux`
 
 ## CUDA x86_64-linux
 
-- `torch213-cxx11-cu126-x86_64-linux`
-- `torch213-cxx11-cu130-x86_64-linux`
-- `torch213-cxx11-cu132-x86_64-linux`
-- `torch214-cxx11-cu126-x86_64-linux`
-- `torch214-cxx11-cu130-x86_64-linux`
-- `torch214-cxx11-cu132-x86_64-linux`
+- `torch213-cu126-x86_64-linux`
+- `torch213-cu130-x86_64-linux`
+- `torch213-cu132-x86_64-linux`
+- `torch214-cu126-x86_64-linux`
+- `torch214-cu130-x86_64-linux`
+- `torch214-cu132-x86_64-linux`
 
 ## ROCm x86_64-linux
 
-- `torch213-cxx11-rocm71-x86_64-linux`
-- `torch213-cxx11-rocm72-x86_64-linux`
-- `torch214-cxx11-rocm714-x86_64-linux`
-- `torch214-cxx11-rocm72-x86_64-linux`
+- `torch213-rocm71-x86_64-linux`
+- `torch213-rocm72-x86_64-linux`
+- `torch214-rocm714-x86_64-linux`
+- `torch214-rocm72-x86_64-linux`
 
 ## XPU x86_64-linux
 
-- `torch213-cxx11-xpu20260-x86_64-linux`
-- `torch214-cxx11-xpu20261-x86_64-linux`
+- `torch213-xpu20260-x86_64-linux`
+- `torch214-xpu20261-x86_64-linux`
 
 ## Python-only kernels
 

diff --git a/docs/source/builder/build.md b/docs/source/builder/build.md
@@ -98,7 +98,7 @@ using:
 
 ```bash
 $ rm -rf .venv  # Remove existing venv if any.
-$ kernel-builder devshell --variant torch212-cxx11-rocm71-x86_64-linux
+$ kernel-builder devshell --variant torch214-rocm72-x86_64-linux
 ```
 
 For an editor-driven workflow with `direnv` activating the devshell on

diff --git a/docs/source/builder/ide-setup.md b/docs/source/builder/ide-setup.md
@@ -88,7 +88,7 @@ the `Creating new venv environment in path: './.venv'` line from the
 To pin a non-default build variant, name it explicitly:
 
 ```bash
-$ echo 'use flake .#devShells.torch212-cxx11-rocm71-x86_64-linux' > .envrc
+$ echo 'use flake .#devShells.torch214-rocm71-x86_64-linux' > .envrc
 $ direnv allow
 ```
 
@@ -182,13 +182,13 @@ variant. For example:
 
 ```bash
 # CUDA 13.0
-use flake .#devShells.torch212-cxx11-cu130-x86_64-linux
+use flake .#devShells.torch214-cu130-x86_64-linux
 
 # ROCm 7.1
-use flake .#devShells.torch212-cxx11-rocm71-x86_64-linux
+use flake .#devShells.torch214-rocm72-x86_64-linux
 
 # XPU
-use flake .#devShells.torch212-cxx11-xpu20253-x86_64-linux
+use flake .#devShells.torch212-xpu20261-x86_64-linux
 ```
 
 Remove `.venv/` first if it was created against a different variant,

diff --git a/docs/source/builder/writing-kernels.md b/docs/source/builder/writing-kernels.md
@@ -539,7 +539,7 @@ ROCm, XPU, Metal, CPU. If you would like to the tests for a specific build
 variant, you can use `nix run .#ciTests.<variant>`. For instance:
 
 ```bash
-$ nix run .#ciTests.torch210-cxx11-cpu-x86_64-linux
+$ nix run .#ciTests.torch210-cpu-x86_64-linux
 ```
 
 When running the tests on a non-NixOS systems, make sure that

diff --git a/docs/source/kernel-requirements.md b/docs/source/kernel-requirements.md
@@ -46,8 +46,8 @@ displays a corresponding badge in the UI.
 A kernel repository on the Hub must contain a `build` directory. This
 directory contains build variants of a kernel in the form of directories
 following the template
-`<framework><version>-cxx<abiver>-<cu><cudaver>-<arch>-<os>`.
-For example `build/torch26-cxx98-cu118-x86_64-linux`.
+`<framework><version>-<cu><cudaver>-<arch>-<os>`.
+For example `build/torch214-cu130-x86_64-linux`.
 
 The kernel is in the build variant directory and must contain a
 `__init__.py` file. For compatibility with older versions of the

diff --git a/examples/kernels/flake.nix b/examples/kernels/flake.nix
@@ -38,14 +38,13 @@
         {
           name = "cpp20-symbols-kernel";
           path = ./cpp20-symbols;
-          drv = sys: out: out.packages.${sys}.redistributable.${"torch${torchVersion}-cxx11-cpu-${sys}"};
+          drv = sys: out: out.packages.${sys}.redistributable.${"torch${torchVersion}-cpu-${sys}"};
         }
         # This test should check the capabilities of the oldest supported CUDA.
         {
           name = "relu-kernel";
           path = ./relu;
-          drv =
-            sys: out: out.packages.${sys}.redistributable.${"torch${torchVersion}-cxx11-${cudaVersion}-${sys}"};
+          drv = sys: out: out.packages.${sys}.redistributable.${"torch${torchVersion}-${cudaVersion}-${sys}"};
           checkCudaCapabilities = [
             "7.0"
             "7.2"
@@ -62,8 +61,7 @@
           name = "relu-kernel-cu13";
           path = ./relu;
           drv =
-            sys: out:
-            out.packages.${sys}.redistributable.${"torch${torchVersion}-cxx11-${cuda13Version}-${sys}"};
+            sys: out: out.packages.${sys}.redistributable.${"torch${torchVersion}-${cuda13Version}-${sys}"};
           checkCudaCapabilities = [
             "7.5"
             "8.0"
@@ -82,8 +80,7 @@
           # Check arch intersection. 5.0 is dropped because it is not supported.
           name = "relu-archs-subset";
           path = ./relu-archs-subset;
-          drv =
-            sys: out: out.packages.${sys}.redistributable.${"torch${torchVersion}-cxx11-${cudaVersion}-${sys}"};
+          drv = sys: out: out.packages.${sys}.redistributable.${"torch${torchVersion}-${cudaVersion}-${sys}"};
           checkCudaCapabilities = [
             "7.5"
             "8.0"
@@ -110,13 +107,12 @@
         {
           name = "relu-kernel-cpu";
           path = ./relu;
-          drv = sys: out: out.packages.${sys}.redistributable.${"torch${torchVersion}-cxx11-cpu-${sys}"};
+          drv = sys: out: out.packages.${sys}.redistributable.${"torch${torchVersion}-cpu-${sys}"};
         }
         {
           name = "cutlass-gemm-kernel";
           path = ./cutlass-gemm;
-          drv =
-            sys: out: out.packages.${sys}.redistributable.${"torch${torchVersion}-cxx11-${cudaVersion}-${sys}"};
+          drv = sys: out: out.packages.${sys}.redistributable.${"torch${torchVersion}-${cudaVersion}-${sys}"};
         }
         {
           name = "cutlass-gemm-tvm-ffi-kernel";
@@ -127,8 +123,7 @@
         {
           name = "relu-backprop-compile-kernel";
           path = ./relu-backprop-compile;
-          drv =
-            sys: out: out.packages.${sys}.redistributable.${"torch${torchVersion}-cxx11-${cudaVersion}-${sys}"};
+          drv = sys: out: out.packages.${sys}.redistributable.${"torch${torchVersion}-${cudaVersion}-${sys}"};
         }
         {
           name = "silu-and-mul-kernel";
@@ -155,14 +150,12 @@
         {
           name = "relu-compiler-flags";
           path = ./relu-compiler-flags;
-          drv =
-            sys: out: out.packages.${sys}.redistributable.${"torch${torchVersion}-cxx11-${cudaVersion}-${sys}"};
+          drv = sys: out: out.packages.${sys}.redistributable.${"torch${torchVersion}-${cudaVersion}-${sys}"};
         }
         {
           name = "relu-invalid-capability";
           path = ./relu-invalid-capability;
-          drv =
-            sys: out: out.packages.${sys}.redistributable.${"torch${torchVersion}-cxx11-${cudaVersion}-${sys}"};
+          drv = sys: out: out.packages.${sys}.redistributable.${"torch${torchVersion}-${cudaVersion}-${sys}"};
           assertFail = true;
           assertFailLogs = [ "empty set of capabilities" ];
         }
@@ -223,16 +216,14 @@
         {
           name = "relu-invalid-capability";
           path = ./relu-invalid-capability;
-          drv =
-            sys: out: out.packages.${sys}.redistributable.${"torch${torchVersion}-cxx11-${rocmVersion}-${sys}"};
+          drv = sys: out: out.packages.${sys}.redistributable.${"torch${torchVersion}-${rocmVersion}-${sys}"};
           assertFail = true;
           assertFailLogs = [ "empty set of architectures" ];
         }
         {
           name = "relu-kernel";
           path = ./relu;
-          drv =
-            sys: out: out.packages.${sys}.redistributable.${"torch${torchVersion}-cxx11-${rocmVersion}-${sys}"};
+          drv = sys: out: out.packages.${sys}.redistributable.${"torch${torchVersion}-${rocmVersion}-${sys}"};
           checkRocmArchs = [
             "gfx906"
             "gfx908"
@@ -254,8 +245,7 @@
           # Check arch intersection. gfx940 is dropped because it is not supported.
           name = "relu-archs-subset";
           path = ./relu-archs-subset;
-          drv =
-            sys: out: out.packages.${sys}.redistributable.${"torch${torchVersion}-cxx11-${rocmVersion}-${sys}"};
+          drv = sys: out: out.packages.${sys}.redistributable.${"torch${torchVersion}-${rocmVersion}-${sys}"};
           checkRocmArchs = [
             "gfx90a"
             "gfx942"
@@ -265,8 +255,7 @@
         {
           name = "relu-compiler-flags";
           path = ./relu-compiler-flags;
-          drv =
-            sys: out: out.packages.${sys}.redistributable.${"torch${torchVersion}-cxx11-${rocmVersion}-${sys}"};
+          drv = sys: out: out.packages.${sys}.redistributable.${"torch${torchVersion}-${rocmVersion}-${sys}"};
         }
       ];
 
@@ -313,8 +302,7 @@
         {
           name = "relu-kernel";
           path = ./relu;
-          drv =
-            sys: out: out.packages.${sys}.redistributable.${"torch${torchVersion}-cxx11-${xpuVersion}-${sys}"};
+          drv = sys: out: out.packages.${sys}.redistributable.${"torch${torchVersion}-${xpuVersion}-${sys}"};
         }
         {
           name = "relu-tvm-ffi-kernel";
@@ -331,14 +319,12 @@
         {
           name = "relu-compiler-flags";
           path = ./relu-compiler-flags;
-          drv =
-            sys: out: out.packages.${sys}.redistributable.${"torch${torchVersion}-cxx11-${xpuVersion}-${sys}"};
+          drv = sys: out: out.packages.${sys}.redistributable.${"torch${torchVersion}-${xpuVersion}-${sys}"};
         }
         {
           name = "cutlass-gemm-kernel";
           path = ./cutlass-gemm;
-          drv =
-            sys: out: out.packages.${sys}.redistributable.${"torch${torchVersion}-cxx11-${xpuVersion}-${sys}"};
+          drv = sys: out: out.packages.${sys}.redistributable.${"torch${torchVersion}-${xpuVersion}-${sys}"};
         }
       ];
 
@@ -365,7 +351,7 @@
           drv =
             sys: _out:
             let
-              variant = "torch${torchVersion}-cxx11-cpu-${sys}";
+              variant = "torch${torchVersion}-cpu-${sys}";
               conflictsFlake = mkKernelOutputs { path = ./symbol-conflicts; };
               conflicts2Flake = mkKernelOutputs { path = ./symbol-conflicts2; };
               conflicts = conflictsFlake.packages.${sys}.redistributable.${variant};

diff --git a/kernel-builder/src/pyproject/templates/torch/build-variants.cmake b/kernel-builder/src/pyproject/templates/torch/build-variants.cmake
@@ -1,5 +1,5 @@
 # Generate a standardized build variant name following the pattern:
-# torch<VERSION>-[cxx11-]<COMPUTE>-<ARCH>-<OS>
+# torch<VERSION>-<COMPUTE>-<ARCH>-<OS>
 # or, when compiled against the Torch stable ABI:
 # torch-stable-abi<VERSION>-<COMPUTE>-<ARCH>-<OS>
 #
@@ -13,9 +13,9 @@
 #   TORCH_STABLE_ABI - Stable ABI version the extension was compiled against (e.g., "2.11");
 #                      when set, TORCH_VERSION is ignored and the prefix becomes
 #                      torch-stable-abi<VERSION> (e.g., "2.11" -> "torch-stable-abi211")
-# Example output: torch27-cxx11-cu124-x86_64-linux (Linux)
-#                 torch27-cu124-x86_64-windows (Windows)
-#                 torch27-metal-aarch64-darwin (macOS)
+# Example output: torch214-cu124-x86_64-linux (Linux)
+#                 torch214-cu124-x86_64-windows (Windows)
+#                 torch214-metal-aarch64-darwin (macOS)
 #                 torch-stable-abi211-cu124-x86_64-linux (Linux, stable ABI)
 #
 function(generate_build_name OUT_BUILD_NAME TORCH_VERSION COMPUTE_FRAMEWORK COMPUTE_VERSION)
@@ -117,12 +117,7 @@ function(generate_build_name OUT_BUILD_NAME TORCH_VERSION COMPUTE_FRAMEWORK COMP
     set(ARCH_OS_STRING "${CPU_ARCH}-${OS_NAME}")
 
     # Assemble the final build name
-    # For non-stable-ABI Linux builds, include cxx11 ABI indicator for compatibility
-    if(NOT ARG_TORCH_STABLE_ABI AND ARCH_OS_STRING MATCHES "-linux$")
-        set(BUILD_NAME "${TORCH_PREFIX}-cxx11-${COMPUTE_STRING}-${ARCH_OS_STRING}")
-    else()
-        set(BUILD_NAME "${TORCH_PREFIX}-${COMPUTE_STRING}-${ARCH_OS_STRING}")
-    endif()
+    set(BUILD_NAME "${TORCH_PREFIX}-${COMPUTE_STRING}-${ARCH_OS_STRING}")
 
     set(${OUT_BUILD_NAME} "${BUILD_NAME}" PARENT_SCOPE)
     message(STATUS "Generated build name: ${BUILD_NAME}")
@@ -136,7 +131,7 @@ endfunction()
 # Arguments:
 #   TARGET_NAME - Name of the target to create the install rule for
 #   PACKAGE_NAME - Python package name (e.g., "activation")
-#   BUILD_VARIANT_NAME - Build variant name (e.g., "torch271-cxx11-cu124-x86_64-linux")
+#   BUILD_VARIANT_NAME - Build variant name (e.g., "torch214-cu124-x86_64-linux")
 #   INSTALL_PREFIX - Base installation directory (defaults to CMAKE_INSTALL_PREFIX)
 #   GPU_ARCHS - List of GPU architectures that were compiled
 #               (optional; when provided for CUDA/ROCm, metadata.json will include
@@ -234,7 +229,7 @@ endfunction()
 # Arguments:
 #   TARGET_NAME - Name of the target to create the install rule for
 #   PACKAGE_NAME - Python package name (e.g., "activation")
-#   BUILD_VARIANT_NAME - Build variant name (e.g., "torch271-cxx11-cu124-x86_64-linux")
+#   BUILD_VARIANT_NAME - Build variant name (e.g., "torch214-cu124-x86_64-linux")
 #   GPU_ARCHS - List of GPU architectures that were compiled
 #               (optional; when provided for CUDA/ROCm, metadata.json will include
 #               a "backend" key with the type and arch list)