diff --git a/circle-mlir/circle-mlir/tools-test/CMakeLists.txt b/circle-mlir/circle-mlir/tools-test/CMakeLists.txt
index 814d9bef556..d21e74e6669 100644
--- a/circle-mlir/circle-mlir/tools-test/CMakeLists.txt
+++ b/circle-mlir/circle-mlir/tools-test/CMakeLists.txt
@@ -1,2 +1,4 @@
 # check gtest is OK
 add_subdirectory(check-gtest)
+# generate onnx models
+add_subdirectory(gen-onnx)
diff --git a/circle-mlir/circle-mlir/tools-test/gen-onnx/CMakeLists.txt b/circle-mlir/circle-mlir/tools-test/gen-onnx/CMakeLists.txt
new file mode 100644
index 00000000000..1da9f83322e
--- /dev/null
+++ b/circle-mlir/circle-mlir/tools-test/gen-onnx/CMakeLists.txt
@@ -0,0 +1,85 @@
+# python3 venv folder
+# NOTE Docker image for CI doesn't use venv
+set(VENV_PATH "${CMAKE_SOURCE_DIR}/infra/overlay/venv")
+
+# Copy test scripts
+unset(GEN_SCRIPT_DEPS )
+macro(COPY_SCRIPT FILENAME)
+  set(SCRIPT_SRC "${CMAKE_CURRENT_SOURCE_DIR}/${FILENAME}")
+  set(SCRIPT_DST "${CMAKE_CURRENT_BINARY_DIR}/${FILENAME}")
+  add_custom_command(
+    OUTPUT ${SCRIPT_DST}
+    COMMAND ${CMAKE_COMMAND} -E copy "${SCRIPT_SRC}" "${SCRIPT_DST}"
+    DEPENDS ${SCRIPT_SRC}
+    COMMENT "gen-onnx: prepare ${FILENAME}"
+  )
+  list(APPEND GEN_SCRIPT_DEPS "${SCRIPT_DST}")
+endmacro(COPY_SCRIPT)
+
+COPY_SCRIPT(run_gen_onnx.sh)
+COPY_SCRIPT(run_gen_onnx.py)
+
+# Models folder for unit testing
+set(PYTORCH_UNIT_PATH "${CMAKE_SOURCE_DIR}/models/unit")
+set(PYTORCH_NET_PATH "${CMAKE_SOURCE_DIR}/models/net")
+
+# Pytorch script files
+unset(PYTORCH_UNIT_ITEMS )
+# gather all the names in /models/unit
+file(GLOB PYTORCH_MODELS_ITEMS RELATIVE ${PYTORCH_UNIT_PATH} ${PYTORCH_UNIT_PATH}/*)
+# for each name, if the item is a folder, add to the list
+foreach(PYTORCH_ITEM IN ITEMS ${PYTORCH_MODELS_ITEMS})
+  if(IS_DIRECTORY ${PYTORCH_UNIT_PATH}/${PYTORCH_ITEM})
+    if(EXISTS ${PYTORCH_UNIT_PATH}/${PYTORCH_ITEM}/__init__.py)
+      list(APPEND PYTORCH_UNIT_ITEMS ${PYTORCH_ITEM})
+    endif()
+  endif()
+endforeach()
+
+unset(PYTORCH_MODELS_ITEMS)
+unset(PYTORCH_NET_ITEMS )
+# gather all the names in /models/net
+file(GLOB PYTORCH_MODELS_ITEMS RELATIVE ${PYTORCH_NET_PATH} ${PYTORCH_NET_PATH}/*)
+# for each name, if the item is a folder, add to the list
+foreach(PYTORCH_ITEM IN ITEMS ${PYTORCH_MODELS_ITEMS})
+  if(IS_DIRECTORY ${PYTORCH_NET_PATH}/${PYTORCH_ITEM})
+    if(EXISTS ${PYTORCH_NET_PATH}/${PYTORCH_ITEM}/__init__.py)
+      list(APPEND PYTORCH_NET_ITEMS ${PYTORCH_ITEM})
+    endif()
+  endif()
+endforeach()
+
+unset(GEN_FILES_DEPS)
+# for each items in the list, run run_gen_onnx to generate onnx model
+foreach(PYTORCH_ITEM IN ITEMS ${PYTORCH_UNIT_ITEMS})
+  set(ONNX_FILE ${PYTORCH_ITEM}.onnx)
+  add_custom_command(OUTPUT ${ONNX_FILE}
+    COMMAND bash run_gen_onnx.sh
+            ${VENV_PATH} ${PYTORCH_UNIT_PATH} ${PYTORCH_ITEM} ${ONNX_FILE}
+    DEPENDS ${CMAKE_CURRENT_BINARY_DIR}/run_gen_onnx.sh
+            ${CMAKE_CURRENT_BINARY_DIR}/run_gen_onnx.py
+            ${PYTORCH_UNIT_PATH}/${PYTORCH_ITEM}/__init__.py
+    WORKING_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR}
+    COMMENT "Generate ${ONNX_FILE}"
+  )
+  list(APPEND GEN_FILES_DEPS ${ONNX_FILE})
+endforeach()
+
+foreach(PYTORCH_ITEM IN ITEMS ${PYTORCH_NET_ITEMS})
+  set(ONNX_FILE ${PYTORCH_ITEM}.onnx)
+  add_custom_command(OUTPUT ${ONNX_FILE}
+    COMMAND bash run_gen_onnx.sh
+            ${VENV_PATH} ${PYTORCH_NET_PATH} ${PYTORCH_ITEM} ${ONNX_FILE}
+    DEPENDS ${CMAKE_CURRENT_BINARY_DIR}/run_gen_onnx.sh
+            ${CMAKE_CURRENT_BINARY_DIR}/run_gen_onnx.py
+            ${PYTORCH_NET_PATH}/${PYTORCH_ITEM}/__init__.py
+    WORKING_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR}
+    COMMENT "Generate ${ONNX_FILE}"
+  )
+  list(APPEND GEN_FILES_DEPS ${ONNX_FILE})
+endforeach()
+
+# gen_onnx_target is used to make build dependency
+add_custom_target(gen_onnx_target ALL DEPENDS ${GEN_SCRIPT_DEPS} ${GEN_FILES_DEPS})
+
+set(GEN_ONNX_PATH ${CMAKE_CURRENT_BINARY_DIR} PARENT_SCOPE)
diff --git a/circle-mlir/circle-mlir/tools-test/gen-onnx/run_gen_onnx.py b/circle-mlir/circle-mlir/tools-test/gen-onnx/run_gen_onnx.py
new file mode 100644
index 00000000000..ef7e071ce86
--- /dev/null
+++ b/circle-mlir/circle-mlir/tools-test/gen-onnx/run_gen_onnx.py
@@ -0,0 +1,64 @@
+# Copyright (c) 2025 Samsung Electronics Co., Ltd. All Rights Reserved
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#    http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import torch
+import importlib
+import sys
+
+from pathlib import Path
+
+
+def generate_onnx(models_root, model_name, onnx_file):
+    sys.path.append(models_root)
+    module = importlib.import_module(model_name)
+
+    # default: refer https://github.com/pytorch/pytorch/blob/master/torch/onnx/utils.py
+    # and https://github.com/pytorch/pytorch/blob/master/torch/onnx/_constants.py
+    # and https://github.com/pytorch/pytorch/blob/master/tools/onnx/update_default_opset_version.py
+    opset_version = 14
+    if hasattr(module._model_, 'onnx_opset_version'):
+        opset_version = module._model_.onnx_opset_version()
+
+    m_keys = module.__dict__.keys()
+
+    if '_io_names_' in m_keys and '_dynamic_axes_' in m_keys:
+        # refer https://github.com/onnx/onnx/issues/654#issuecomment-521233285
+        # purpose of this is to set dynamic shape for inputs or inputs
+        # magic(?) is to set input/output names, and then set dyanmic shape by name/dim
+        # example) set output dim(0) as unknown
+        #    _io_names_ = [['input'], ['output']]
+        #    _dynamic_axes_ = {'output': {0: '?'}}
+        torch.onnx.export(module._model_,
+                          module._inputs_,
+                          onnx_file,
+                          input_names=module._io_names_[0],
+                          output_names=module._io_names_[1],
+                          dynamic_axes=module._dynamic_axes_,
+                          opset_version=opset_version)
+    else:
+        torch.onnx.export(module._model_,
+                          module._inputs_,
+                          onnx_file,
+                          opset_version=opset_version)
+
+    if hasattr(module._model_, 'post_process'):
+        module._model_.post_process(onnx_file)
+
+
+if __name__ == "__main__":
+    if len(sys.argv) != 4:
+        thispath = Path(sys.argv[0])
+        sys.exit("Usage: " + thispath.name + " [models_root] [model_name] [onnx_file]")
+
+    generate_onnx(sys.argv[1], sys.argv[2], sys.argv[3])
diff --git a/circle-mlir/circle-mlir/tools-test/gen-onnx/run_gen_onnx.sh b/circle-mlir/circle-mlir/tools-test/gen-onnx/run_gen_onnx.sh
new file mode 100644
index 00000000000..434d45a25b4
--- /dev/null
+++ b/circle-mlir/circle-mlir/tools-test/gen-onnx/run_gen_onnx.sh
@@ -0,0 +1,48 @@
+#!/bin/bash
+
+# This script executes run_gen_onnx.py file to generate ONNX model
+#
+# HOW TO USE
+#
+# ./run_gen_onnx.sh <path/to/venv_dir> <path/to/models> <model_name> <onnx_name>
+# venv_dir   : python virtual environment home directory
+# models     : path where python modules exist
+# model_name : name of model
+# onnx_name  : name of onnx file
+
+THIS_SCRIPT_PATH="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+PY_SCRIPT_PATH="${THIS_SCRIPT_PATH}/run_gen_onnx.py"
+
+VENV_PATH="$1"; shift
+MODELS_ROOT_PATH="$1"; shift
+MODEL_NAME="$1"; shift
+ONNX_NAME="$1"; shift
+
+PASSED_TAG="${ONNX_NAME}.passed"
+GENERATE_LOG="${ONNX_NAME}.log"
+rm -f "${PASSED_TAG}"
+
+cat > "${GENERATE_LOG}" <(
+  exec 2>&1
+  set -ex
+
+  # NOTE enter venv if exist
+  if [[ -f "${VENV_PATH}/bin/activate" ]]; then
+    source "${VENV_PATH}/bin/activate"
+  fi
+
+  "python3" "${PY_SCRIPT_PATH}" "${MODELS_ROOT_PATH}" "${MODEL_NAME}" "${ONNX_NAME}"
+  if [[ $? -eq 0 ]]; then
+    touch "${PASSED_TAG}"
+  fi
+
+  if [[ -f "${VENV_PATH}/bin/activate" ]]; then
+    deactivate
+  fi
+)
+
+if [[ ! -f "${PASSED_TAG}" ]]; then
+  exit 255
+fi
+rm -f "${PASSED_TAG}"
+exit 0
diff --git a/circle-mlir/models/unit/Add_F32_R4/__init__.py b/circle-mlir/models/unit/Add_F32_R4/__init__.py
new file mode 100644
index 00000000000..f58d7e4ab8f
--- /dev/null
+++ b/circle-mlir/models/unit/Add_F32_R4/__init__.py
@@ -0,0 +1,19 @@
+import torch
+
+
+# Generate Add operator with Float32, Rank-4
+class net_add(torch.nn.Module):
+    def __init__(self):
+        super().__init__()
+
+    def forward(self, inputs):
+        return torch.add(inputs[0], inputs[1])
+
+    def onnx_opset_version(self):
+        # TODO set version
+        return 10
+
+
+_model_ = net_add()
+
+_inputs_ = [torch.randn(1, 2, 3, 3), torch.randn(1, 2, 3, 3)]
diff --git a/circle-mlir/models/unit/Add_F32_R4_C1/__init__.py b/circle-mlir/models/unit/Add_F32_R4_C1/__init__.py
new file mode 100644
index 00000000000..55bb946c2ad
--- /dev/null
+++ b/circle-mlir/models/unit/Add_F32_R4_C1/__init__.py
@@ -0,0 +1,22 @@
+import torch
+import numpy as np
+
+
+# Generate Add operator with Float32, Rank-4 with Constant input
+class net_add(torch.nn.Module):
+    def __init__(self):
+        super().__init__()
+        rng = np.random.default_rng(seed=123)
+        self.C1 = torch.from_numpy(rng.random((1, 2, 3, 3), dtype=np.float32))
+
+    def forward(self, inputs):
+        return torch.add(inputs[0], self.C1)
+
+    def onnx_opset_version(self):
+        # TODO set version
+        return 10
+
+
+_model_ = net_add()
+
+_inputs_ = [torch.randn(1, 2, 3, 3)]