diff --git a/src/frontends/tensorflow/src/op/block_lstm.cpp b/src/frontends/tensorflow/src/op/block_lstm.cpp
index 925906cdaca..c9e9929fbc7 100644
--- a/src/frontends/tensorflow/src/op/block_lstm.cpp
+++ b/src/frontends/tensorflow/src/op/block_lstm.cpp
@@ -5,7 +5,6 @@
 #include "helper_ops/block_lstm.hpp"
 
 #include "common_op_table.hpp"
-#include "ngraph/validation_util.hpp"
 #include "openvino/core/validation_util.hpp"
 #include "openvino/frontend/tensorflow/node_context.hpp"
 
diff --git a/src/frontends/tensorflow/src/op/sparse_reshape.cpp b/src/frontends/tensorflow/src/op/sparse_reshape.cpp
index 6a09ecb2872..10f830cf4b2 100644
--- a/src/frontends/tensorflow/src/op/sparse_reshape.cpp
+++ b/src/frontends/tensorflow/src/op/sparse_reshape.cpp
@@ -5,14 +5,12 @@
 #include "common_op_table.hpp"
 #include "helper_ops/sparse_fill_empty_rows.hpp"
 #include "helper_ops/sparse_segment_ops.hpp"
-#include "ngraph/validation_util.hpp"
 #include "openvino/core/validation_util.hpp"
 #include "openvino/frontend/tensorflow/node_context.hpp"
 #include "openvino/opsets/opset8.hpp"
 
 using namespace std;
 using namespace ov::opset8;
-using namespace ngraph;
 
 namespace ov {
 namespace frontend {
diff --git a/src/frontends/tensorflow/tests/basic_api.cpp b/src/frontends/tensorflow/tests/basic_api.cpp
index a3f95fc5a87..301f21f9e04 100644
--- a/src/frontends/tensorflow/tests/basic_api.cpp
+++ b/src/frontends/tensorflow/tests/basic_api.cpp
@@ -6,7 +6,6 @@
 
 #include "tf_utils.hpp"
 
-using namespace ngraph;
 using namespace ov::frontend;
 using namespace ov::frontend::tensorflow::tests;
 
diff --git a/src/frontends/tensorflow/tests/convert_model.cpp b/src/frontends/tensorflow/tests/convert_model.cpp
index 22f46a3a94d..fc00a678496 100644
--- a/src/frontends/tensorflow/tests/convert_model.cpp
+++ b/src/frontends/tensorflow/tests/convert_model.cpp
@@ -6,7 +6,6 @@
 
 #include "tf_utils.hpp"
 
-using namespace ngraph;
 using namespace ov::frontend;
 using namespace ov::frontend::tensorflow::tests;
 
diff --git a/src/frontends/tensorflow/tests/convert_saved_model.cpp b/src/frontends/tensorflow/tests/convert_saved_model.cpp
index d0ce96deefc..b8f62a1c004 100644
--- a/src/frontends/tensorflow/tests/convert_saved_model.cpp
+++ b/src/frontends/tensorflow/tests/convert_saved_model.cpp
@@ -4,7 +4,7 @@
 
 #include <openvino/opsets/opset10.hpp>
 
-#include "common_test_utils/ngraph_test_utils.hpp"
+#include "conversion_with_reference.hpp"
 #include "gtest/gtest.h"
 #include "test_common.hpp"
 #include "tf_utils.hpp"
@@ -14,7 +14,7 @@ using namespace ov;
 using namespace ov::opset10;
 using namespace ov::frontend::tensorflow::tests;
 
-TEST_F(TransformationTestsF, SavedModelProgramOnly) {
+TEST_F(FrontEndConversionWithReferenceTestsF, SavedModelProgramOnly) {
     {
         model = convert_model("saved_model_program-only");
 
@@ -40,7 +40,7 @@ TEST_F(TransformationTestsF, SavedModelProgramOnly) {
     }
 }
 
-TEST_F(TransformationTestsF, SavedModelVariables) {
+TEST_F(FrontEndConversionWithReferenceTestsF, SavedModelVariables) {
     { model = convert_model("saved_model_variables"); }
     {
         // create a reference graph
@@ -52,7 +52,7 @@ TEST_F(TransformationTestsF, SavedModelVariables) {
     }
 }
 
-TEST_F(TransformationTestsF, SavedModelWithInputIntegerType) {
+TEST_F(FrontEndConversionWithReferenceTestsF, SavedModelWithInputIntegerType) {
     {
         model = convert_model("saved_model_with_gather",
                               nullptr,
@@ -89,7 +89,7 @@ TEST_F(TransformationTestsF, SavedModelWithInputIntegerType) {
     }
 }
 
-TEST_F(TransformationTestsF, SavedModelMultipleTensorNames) {
+TEST_F(FrontEndConversionWithReferenceTestsF, SavedModelMultipleTensorNames) {
     // The test aims to check tensor names of input and output tensors
     // it checks that TF FE preserved user specific names for input and output tensor
     // and exclude internal names
diff --git a/src/frontends/tensorflow/tests/convert_tricky_models.cpp b/src/frontends/tensorflow/tests/convert_tricky_models.cpp
index 470cc7bc734..06be394b9bf 100644
--- a/src/frontends/tensorflow/tests/convert_tricky_models.cpp
+++ b/src/frontends/tensorflow/tests/convert_tricky_models.cpp
@@ -8,7 +8,7 @@
 #include <openvino/opsets/opset10.hpp>
 #include <transformations/common_optimizations/moc_transformations.hpp>
 
-#include "common_test_utils/ngraph_test_utils.hpp"
+#include "conversion_with_reference.hpp"
 #include "gtest/gtest.h"
 #include "test_common.hpp"
 #include "tf_utils.hpp"
@@ -107,7 +107,7 @@ TEST(FrontEndConvertTrickyModels, model_with_output_shapes) {
     }
 }
 
-TEST_F(TransformationTestsF, AssertAndStringTensors) {
+TEST_F(FrontEndConversionWithReferenceTestsF, AssertAndStringTensors) {
     {
         model = convert_model("string_tensors_model/string_tensors_model.pbtxt");
         // TODO: investigate - why we have redundant nodes after the conversion
@@ -123,12 +123,12 @@ TEST_F(TransformationTestsF, AssertAndStringTensors) {
     }
 }
 
-TEST_F(TransformationTestsF, UnsortedNodes) {
+TEST_F(FrontEndConversionWithReferenceTestsF, UnsortedNodes) {
     { model = convert_model("forward_edge_model_unsorted/forward_edge_model_unsorted.pbtxt"); }
     { model_ref = convert_model("forward_edge_model/forward_edge_model.pbtxt"); }
 }
 
-TEST_F(TransformationTestsF, ModelWithSwishF32BodyGraph) {
+TEST_F(FrontEndConversionWithReferenceTestsF, ModelWithSwishF32BodyGraph) {
     {
         model = convert_model("swish_f32/swish_f32.pbtxt");
         // need to call shape inference since body graphs can be injected with undefined shapes
@@ -146,7 +146,7 @@ TEST_F(TransformationTestsF, ModelWithSwishF32BodyGraph) {
     }
 }
 
-TEST_F(TransformationTestsF, PartitionedCall) {
+TEST_F(FrontEndConversionWithReferenceTestsF, PartitionedCall) {
     {
         model = convert_model("partitioned_call/partitioned_call.pbtxt");
         // need to call shape inference since body graphs can be injected with undefined shapes
@@ -163,7 +163,7 @@ TEST_F(TransformationTestsF, PartitionedCall) {
     }
 }
 
-TEST_F(TransformationTestsF, ModelWithIf) {
+TEST_F(FrontEndConversionWithReferenceTestsF, ModelWithIf) {
     { model = convert_model("model_with_if/model_with_if.pbtxt"); }
     {
         // create then branch body graph
@@ -196,7 +196,7 @@ TEST_F(TransformationTestsF, ModelWithIf) {
     }
 }
 
-TEST_F(TransformationTestsF, InjectedBodyAndIf) {
+TEST_F(FrontEndConversionWithReferenceTestsF, InjectedBodyAndIf) {
     {
         model = convert_model("injected_body_and_if/injected_body_and_if.pbtxt");
         // need to call shape inference since body graphs can be injected with undefined shapes
@@ -235,11 +235,15 @@ TEST_F(TransformationTestsF, InjectedBodyAndIf) {
     }
 }
 
-TEST_F(TransformationTestsF, ModelWithDilatedGroupConvolution) {
+TEST_F(FrontEndConversionWithReferenceTestsF, ModelWithDilatedGroupConvolution) {
     {
         model = convert_model("dilated_gconv_model/dilated_gconv_model.pbtxt");
         // need to call MOC to fuse BatchToSpace/SpaceToBatch with GroupConvolution
         manager.register_pass<pass::MOCTransformations>(false);
+        // TODO: enable ATTRIBUTES, CONST_VALUES and ACCURACY checks, CVS-111900
+        comparator.disable(FunctionsComparator::CmpValues::ATTRIBUTES);
+        comparator.disable(FunctionsComparator::CmpValues::CONST_VALUES);
+        comparator.disable(FunctionsComparator::CmpValues::ACCURACY);
     }
     {
         auto x = make_shared<Parameter>(f32, Shape{1, 129, 257, 384});
@@ -259,7 +263,7 @@ TEST_F(TransformationTestsF, ModelWithDilatedGroupConvolution) {
     }
 }
 
-TEST_F(TransformationTestsF, ModelWithSaveV2) {
+TEST_F(FrontEndConversionWithReferenceTestsF, ModelWithSaveV2) {
     {
         model = convert_model("model_savev2/model_savev2.pbtxt");
         // need to call shape inference since body graphs can be injected with undefined shapes
@@ -275,7 +279,7 @@ TEST_F(TransformationTestsF, ModelWithSaveV2) {
     }
 }
 
-TEST_F(TransformationTestsF, ModelWithConstResultSubgraphs) {
+TEST_F(FrontEndConversionWithReferenceTestsF, ModelWithConstResultSubgraphs) {
     { model = convert_model("model_with_const_result/model_with_const_result.pbtxt"); }
     {
         // create a reference graph
@@ -298,7 +302,7 @@ TEST_F(TransformationTestsF, ModelWithConstResultSubgraphs) {
     }
 }
 
-TEST_F(TransformationTestsF, ModelWithIteratorGetNext) {
+TEST_F(FrontEndConversionWithReferenceTestsF, ModelWithIteratorGetNext) {
     { model = convert_model("model_with_iterator_get_next/model_with_iterator_get_next.pbtxt"); }
     {
         // create a reference graph
@@ -310,7 +314,7 @@ TEST_F(TransformationTestsF, ModelWithIteratorGetNext) {
     }
 }
 
-TEST_F(TransformationTestsF, ModelWithQueueOperations) {
+TEST_F(FrontEndConversionWithReferenceTestsF, ModelWithQueueOperations) {
     { model = convert_model("model_with_queue_ops/model_with_queue_ops.pbtxt"); }
     {
         // create a reference graph
@@ -322,7 +326,7 @@ TEST_F(TransformationTestsF, ModelWithQueueOperations) {
     }
 }
 
-TEST_F(TransformationTestsF, ModelWithQueueOperations2) {
+TEST_F(FrontEndConversionWithReferenceTestsF, ModelWithQueueOperations2) {
     { model = convert_model("model_with_queue_ops2/model_with_queue_ops2.pbtxt"); }
     {
         // create a reference graph
@@ -336,7 +340,7 @@ TEST_F(TransformationTestsF, ModelWithQueueOperations2) {
     }
 }
 
-TEST_F(TransformationTestsF, ModelWithLookupTableOperations) {
+TEST_F(FrontEndConversionWithReferenceTestsF, ModelWithLookupTableOperations) {
     { model = convert_model("model_with_lookup_table/model_with_lookup_table.pbtxt"); }
     {
         // create a reference graph
@@ -348,7 +352,7 @@ TEST_F(TransformationTestsF, ModelWithLookupTableOperations) {
     }
 }
 
-TEST_F(TransformationTestsF, ModelWithIteratorGetNextAndUnsupportedOp) {
+TEST_F(FrontEndConversionWithReferenceTestsF, ModelWithIteratorGetNextAndUnsupportedOp) {
     { model = convert_model("unsupported_op_itergetnext/unsupported_op_itergetnext.pb"); }
     {
         // create then branch body graph
@@ -360,8 +364,12 @@ TEST_F(TransformationTestsF, ModelWithIteratorGetNextAndUnsupportedOp) {
     }
 }
 
-TEST_F(TransformationTestsF, ModelWithMultioutputBodyGraphNode) {
-    { model = convert_model("partitioned_call2/partitioned_call2.pbtxt"); }
+TEST_F(FrontEndConversionWithReferenceTestsF, ModelWithMultioutputBodyGraphNode) {
+    {
+        model = convert_model("partitioned_call2/partitioned_call2.pbtxt");
+        // TODO: enable ATTRIBUTES check, CVS-111901
+        comparator.disable(FunctionsComparator::CmpValues::ATTRIBUTES);
+    }
     {
         auto x = make_shared<Parameter>(i32, Shape{5});
         auto y = make_shared<Parameter>(i32, Shape{5});
@@ -375,7 +383,7 @@ TEST_F(TransformationTestsF, ModelWithMultioutputBodyGraphNode) {
     }
 }
 
-TEST_F(TransformationTestsF, ModelWithEmptyTensorListAndPushBack) {
+TEST_F(FrontEndConversionWithReferenceTestsF, ModelWithEmptyTensorListAndPushBack) {
     { model = convert_model("empty_tensor_list/empty_tensor_list.pb"); }
     {
         auto x = make_shared<Parameter>(f32, Shape{2, 3, 5});
@@ -397,7 +405,7 @@ TEST_F(TransformationTestsF, ModelWithEmptyTensorListAndPushBack) {
     }
 }
 
-TEST_F(TransformationTestsF, ModelWithAssertNode) {
+TEST_F(FrontEndConversionWithReferenceTestsF, ModelWithAssertNode) {
     { model = convert_model("model_with_assert/model_with_assert.pb"); }
     {
         auto x = make_shared<Parameter>(i32, PartialShape{Dimension::dynamic()});
@@ -407,7 +415,7 @@ TEST_F(TransformationTestsF, ModelWithAssertNode) {
     }
 }
 
-TEST_F(TransformationTestsF, PartitionedCallWithUnique) {
+TEST_F(FrontEndConversionWithReferenceTestsF, PartitionedCallWithUnique) {
     // This test aims to test named output ports for Unique operation
     { model = convert_model("partitioned_call_with_unique/partitioned_call_with_unique.pb"); }
     {
@@ -421,7 +429,7 @@ TEST_F(TransformationTestsF, PartitionedCallWithUnique) {
     }
 }
 
-TEST_F(TransformationTestsF, RaggedTensorToSparse) {
+TEST_F(FrontEndConversionWithReferenceTestsF, RaggedTensorToSparse) {
     // This test aims to test named output ports for RaggedTensorToSparse operation
     // also, it tests propagation of custom type (specified in the extension) to Parameter node in the parent graph
     {
@@ -451,7 +459,7 @@ TEST_F(TransformationTestsF, RaggedTensorToSparse) {
     }
 }
 
-TEST_F(TransformationTestsF, MetaGraphVariables) {
+TEST_F(FrontEndConversionWithReferenceTestsF, MetaGraphVariables) {
     {
         model = convert_model("metagraph_variables/graph.meta");
         model->validate_nodes_and_infer_types();
@@ -468,7 +476,7 @@ TEST_F(TransformationTestsF, MetaGraphVariables) {
     }
 }
 
-TEST_F(TransformationTestsF, MetaGraphCut) {
+TEST_F(FrontEndConversionWithReferenceTestsF, MetaGraphCut) {
     {
         model = convert_model("metagraph_variables/graph.meta", nullptr, {"y"});
         model->validate_nodes_and_infer_types();
@@ -485,7 +493,7 @@ TEST_F(TransformationTestsF, MetaGraphCut) {
     }
 }
 
-TEST_F(TransformationTestsF, MetaGraphCutInputTensor) {
+TEST_F(FrontEndConversionWithReferenceTestsF, MetaGraphCutInputTensor) {
     {
         model = convert_model("metagraph_variables/graph.meta",
                               nullptr,
@@ -504,7 +512,7 @@ TEST_F(TransformationTestsF, MetaGraphCutInputTensor) {
     }
 }
 
-TEST_F(TransformationTestsF, MetaGraphCutOutputTensor) {
+TEST_F(FrontEndConversionWithReferenceTestsF, MetaGraphCutOutputTensor) {
     {
         model = convert_model("metagraph_variables/graph.meta",
                               nullptr,
@@ -523,7 +531,7 @@ TEST_F(TransformationTestsF, MetaGraphCutOutputTensor) {
     }
 }
 
-TEST_F(TransformationTestsF, MetaGraphCutIdentity) {
+TEST_F(FrontEndConversionWithReferenceTestsF, MetaGraphCutIdentity) {
     {
         model = convert_model("metagraph_variables/graph.meta",
                               nullptr,
@@ -542,7 +550,7 @@ TEST_F(TransformationTestsF, MetaGraphCutIdentity) {
     }
 }
 
-TEST_F(TransformationTestsF, SplitInFunction) {
+TEST_F(FrontEndConversionWithReferenceTestsF, SplitInFunction) {
     {
         // create FAKE conversion extension for Split using named ports, this is not required for Split, but it tests
         // how named ports will work if there is one name and many outputs associated with it
@@ -572,7 +580,7 @@ TEST_F(TransformationTestsF, SplitInFunction) {
     }
 }
 
-TEST_F(TransformationTestsF, ResourceGatherModel) {
+TEST_F(FrontEndConversionWithReferenceTestsF, ResourceGatherModel) {
     // This test aims to check basic support of ResourceGather operation
     // and cutting an input model with specified shapes and types
     {
@@ -600,7 +608,7 @@ TEST_F(TransformationTestsF, ResourceGatherModel) {
     }
 }
 
-TEST_F(TransformationTestsF, NonMaxSuppressionWithNamedOutputs) {
+TEST_F(FrontEndConversionWithReferenceTestsF, NonMaxSuppressionWithNamedOutputs) {
     // The purpose of this test is to check that named output ports of TensorFlow NMS operation are connected correctly
     // to its consumers
     { model = convert_model("nms_named_outputs/nms_named_outputs.pb"); }
@@ -668,7 +676,7 @@ TEST_F(TransformationTestsF, NonMaxSuppressionWithNamedOutputs) {
     }
 }
 
-TEST_F(TransformationTestsF, PartitionedCallsWithConvInBodyGraphs) {
+TEST_F(FrontEndConversionWithReferenceTestsF, PartitionedCallsWithConvInBodyGraphs) {
     // The test aims to check that the conversion for the body graphs is performed with set input shapes
     // that allows to get more optimized ov::Model for the body graphs.
     // In particular, we check that the resulted graph contains Convolution operations instead of GroupConvolution
diff --git a/src/frontends/tensorflow/tests/convert_unsupported.cpp b/src/frontends/tensorflow/tests/convert_unsupported.cpp
index 70a21635429..21d350cc2da 100644
--- a/src/frontends/tensorflow/tests/convert_unsupported.cpp
+++ b/src/frontends/tensorflow/tests/convert_unsupported.cpp
@@ -11,7 +11,7 @@
 #include <openvino/op/util/framework_node.hpp>
 #include <openvino/opsets/opset10.hpp>
 
-#include "common_test_utils/ngraph_test_utils.hpp"
+#include "conversion_with_reference.hpp"
 #include "tf_framework_node.hpp"
 #include "tf_utils.hpp"
 #include "utils.hpp"
@@ -142,7 +142,7 @@ TEST(FrontEndConvertModelTest, test_unsupported_tf1_while) {
     }
 }
 
-TEST_F(TransformationTestsF, ModelWithDynamicType) {
+TEST_F(FrontEndConversionWithReferenceTestsF, ModelWithDynamicType) {
     { model = convert_model_partially("dynamic_type_model/dynamic_type_model.pb"); }
     {
         auto x = make_shared<Parameter>(f32, Shape{2, 3});
diff --git a/src/frontends/tensorflow/tests/telemetry.cpp b/src/frontends/tensorflow/tests/telemetry.cpp
index 3caf1cc01b7..5f9b433a92a 100644
--- a/src/frontends/tensorflow/tests/telemetry.cpp
+++ b/src/frontends/tensorflow/tests/telemetry.cpp
@@ -79,16 +79,16 @@ TEST(TFTelemetryTest, test_nonexistent_add) {
                                                              string("nonexistent_add/nonexistent_add.pb"));
     ASSERT_NO_THROW(inputModel = frontEnd->load(model_filename));
     ASSERT_NE(inputModel, nullptr);
-    shared_ptr<ngraph::Function> function;
+    shared_ptr<ov::Model> model;
 
     try {
-        function = frontEnd->convert(inputModel);
+        model = frontEnd->convert(inputModel);
         FAIL() << "Non-existent operation Adddd must not be supported by TF FE.";
     } catch (const OpConversionFailure& error) {
         string error_message = error.what();
         string ref_message = "Internal error, no translator found for operation(s): Adddd";
         ASSERT_TRUE(error_message.find(ref_message) != string::npos);
-        ASSERT_EQ(function, nullptr);
+        ASSERT_EQ(model, nullptr);
 
         // check telemetry data
         EXPECT_EQ(m_test_telemetry.m_error_cnt, 0);
diff --git a/src/frontends/tensorflow_common/src/helper_transforms/block_lstm_replacer.cpp b/src/frontends/tensorflow_common/src/helper_transforms/block_lstm_replacer.cpp
index 579f6b1f4f9..d2931eef4fc 100644
--- a/src/frontends/tensorflow_common/src/helper_transforms/block_lstm_replacer.cpp
+++ b/src/frontends/tensorflow_common/src/helper_transforms/block_lstm_replacer.cpp
@@ -8,7 +8,6 @@
 #include <vector>
 
 #include "helper_ops/block_lstm.hpp"
-#include "ngraph/rt_info.hpp"
 #include "openvino/opsets/opset9.hpp"
 #include "openvino/pass/pattern/matcher.hpp"
 #include "openvino/pass/pattern/op/or.hpp"
diff --git a/src/frontends/tensorflow_common/src/helper_transforms/gru_block_cell_replacer.cpp b/src/frontends/tensorflow_common/src/helper_transforms/gru_block_cell_replacer.cpp
index ee33aa6d84a..5bb65a1dae9 100644
--- a/src/frontends/tensorflow_common/src/helper_transforms/gru_block_cell_replacer.cpp
+++ b/src/frontends/tensorflow_common/src/helper_transforms/gru_block_cell_replacer.cpp
@@ -8,7 +8,6 @@
 #include <vector>
 
 #include "helper_ops/gru_block_cell.hpp"
-#include "ngraph/rt_info.hpp"
 #include "openvino/opsets/opset9.hpp"
 #include "openvino/pass/pattern/matcher.hpp"
 #include "openvino/pass/pattern/op/or.hpp"
diff --git a/src/frontends/tests/frontend/shared/include/conversion_with_reference.hpp b/src/frontends/tests/frontend/shared/include/conversion_with_reference.hpp
new file mode 100644
index 00000000000..299a07be08e
--- /dev/null
+++ b/src/frontends/tests/frontend/shared/include/conversion_with_reference.hpp
@@ -0,0 +1,25 @@
+// Copyright (C) 2018-2023 Intel Corporation
+// SPDX-License-Identifier: Apache-2.0
+//
+
+#pragma once
+
+#include <gtest/gtest.h>
+
+#include "common_test_utils/graph_comparator.hpp"
+#include "openvino/core/model.hpp"
+#include "openvino/pass/manager.hpp"
+
+// test fixture to check that the resulted graph generated by the frontend matches the reference
+class FrontEndConversionWithReferenceTestsF : virtual public ::testing::Test {
+public:
+    FrontEndConversionWithReferenceTestsF();
+
+    void SetUp() override;
+
+    void TearDown() override;
+
+    std::shared_ptr<ov::Model> model, model_ref;
+    ov::pass::Manager manager;
+    FunctionsComparator comparator;
+};
diff --git a/src/frontends/tests/frontend/shared/src/conversion_with_reference.cpp b/src/frontends/tests/frontend/shared/src/conversion_with_reference.cpp
new file mode 100644
index 00000000000..f321b64fd04
--- /dev/null
+++ b/src/frontends/tests/frontend/shared/src/conversion_with_reference.cpp
@@ -0,0 +1,33 @@
+// Copyright (C) 2018-2023 Intel Corporation
+// SPDX-License-Identifier: Apache-2.0
+//
+
+#include "conversion_with_reference.hpp"
+
+#include "transformations/init_node_info.hpp"
+
+FrontEndConversionWithReferenceTestsF::FrontEndConversionWithReferenceTestsF()
+    : comparator(FunctionsComparator::no_default()) {
+    comparator.enable(FunctionsComparator::CmpValues::NODES);
+    comparator.enable(FunctionsComparator::CmpValues::PRECISIONS);
+    comparator.enable(FunctionsComparator::CmpValues::RUNTIME_KEYS);
+    comparator.enable(FunctionsComparator::CmpValues::SUBGRAPH_DESCRIPTORS);
+    comparator.enable(FunctionsComparator::CmpValues::ATTRIBUTES);
+    comparator.enable(FunctionsComparator::CmpValues::CONST_VALUES);
+    comparator.enable(FunctionsComparator::CmpValues::ACCURACY);
+}
+
+void FrontEndConversionWithReferenceTestsF::SetUp() {
+    manager.register_pass<ov::pass::InitNodeInfo>();
+}
+
+void FrontEndConversionWithReferenceTestsF::TearDown() {
+    OPENVINO_ASSERT(model != nullptr, "Test Model is not initialized.");
+    OPENVINO_ASSERT(model_ref != nullptr, "Reference Test Model is not initialized.");
+
+    manager.run_passes(model);
+    ASSERT_NO_THROW(check_rt_info(model));
+
+    auto res = comparator.compare(model, model_ref);
+    ASSERT_TRUE(res.valid) << res.message;
+}