diff --git a/src/core/shape_inference/include/variadic_split_shape_inference.hpp b/src/core/shape_inference/include/variadic_split_shape_inference.hpp
index 77023a61641725..2694ee5e4594c6 100644
--- a/src/core/shape_inference/include/variadic_split_shape_inference.hpp
+++ b/src/core/shape_inference/include/variadic_split_shape_inference.hpp
@@ -75,7 +75,7 @@ std::vector<TRShape> shape_infer(const VariadicSplit* op,
                         sum_of_splits += (*split_lengths)[i];
                     }
                 }
-                const auto dimension_at_axis = data_shape[axis];
+                const auto& dimension_at_axis = data_shape[axis];
 
                 if (negative_one_idx >= 0 && dimension_at_axis.is_static()) {
                     (*split_lengths)[negative_one_idx] = dimension_at_axis.get_length() - sum_of_splits;
diff --git a/src/core/src/evaluator.hpp b/src/core/src/evaluator.hpp
index 3e9408219b2522..eaa714f53f00d7 100644
--- a/src/core/src/evaluator.hpp
+++ b/src/core/src/evaluator.hpp
@@ -135,7 +135,7 @@ class Evaluator {
             // Request to execute the handleer. Pass what we know about the inputs to the
             // handler and associate the results with the outputs
             std::vector<V> inputs;
-            for (auto v : node->input_values()) {
+            for (const auto& v : node->input_values()) {
                 inputs.push_back(evaluator.get_value_map().at(v));
             }
             std::vector<V> outputs = m_handler(node, inputs);
diff --git a/src/core/src/layout.cpp b/src/core/src/layout.cpp
index 2bec5d0b0e76d4..7c3c29909088fd 100644
--- a/src/core/src/layout.cpp
+++ b/src/core/src/layout.cpp
@@ -448,7 +448,7 @@ std::tuple<PartialShape, Layout> LayoutUtils::find_squeeze(const Layout& src_lay
     res.m_left_size = dst_layout.m_left_size;
     int64_t dst_idx = 0;
     for (int64_t src_idx = 0; src_idx < src_layout.m_left_size; src_idx++) {
-        auto src_dim_name = src_layout.m_index_map.at(src_idx);
+        const auto& src_dim_name = src_layout.m_index_map.at(src_idx);
         if (dst_layout.has_name(src_dim_name)) {
             if (!rank_dynamic) {
                 res_dims[dst_idx] = src_shape[src_idx];
@@ -496,7 +496,7 @@ std::tuple<PartialShape, Layout, size_t> LayoutUtils::find_unsqueeze(const Layou
     res.m_left_size = dst_layout.m_left_size;
     int64_t unset_idx = 0;
     for (auto i = 0; i < dst_layout.m_left_size; i++) {
-        auto dim_name = dst_layout.m_index_map.at(i);
+        const auto& dim_name = dst_layout.m_index_map.at(i);
         if (src_layout.has_name(dim_name)) {
             auto src_idx = src_layout.get_index_by_name(dim_name);
             res.m_names[dim_name] = src_idx + dims_cnt;
diff --git a/src/core/src/op/adaptive_avg_pool.cpp b/src/core/src/op/adaptive_avg_pool.cpp
index cd7ef6f4d32a15..810be9208bb8e3 100644
--- a/src/core/src/op/adaptive_avg_pool.cpp
+++ b/src/core/src/op/adaptive_avg_pool.cpp
@@ -17,8 +17,8 @@ op::v8::AdaptiveAvgPool::AdaptiveAvgPool(const Output<Node>& data, const Output<
 void op::v8::AdaptiveAvgPool::validate_and_infer_types() {
     OV_OP_SCOPE(v8_AdaptiveAvgPool_validate_and_infer_types);
 
-    const auto output_shape = shape_infer(this, ov::util::get_node_input_partial_shapes(*this)).front();
-    set_output_type(0, get_input_element_type(0), output_shape);
+    const auto output_shapes = shape_infer(this, ov::util::get_node_input_partial_shapes(*this));
+    set_output_type(0, get_input_element_type(0), output_shapes[0]);
 }
 
 std::shared_ptr<Node> op::v8::AdaptiveAvgPool::clone_with_new_inputs(const OutputVector& new_args) const {
diff --git a/src/core/src/op/batch_to_space.cpp b/src/core/src/op/batch_to_space.cpp
index a70c2af3a671eb..1d1d8b8b85e2f3 100644
--- a/src/core/src/op/batch_to_space.cpp
+++ b/src/core/src/op/batch_to_space.cpp
@@ -50,8 +50,8 @@ void BatchToSpace::validate_and_infer_types() {
                           "block_shape and crops inputs must have integer element type. Got: ",
                           inputs_integer_et);
 
-    const auto output_shape = shape_infer(this, ov::util::get_node_input_partial_shapes(*this)).front();
-    set_output_type(0, data_et, output_shape);
+    const auto output_shapes = shape_infer(this, ov::util::get_node_input_partial_shapes(*this));
+    set_output_type(0, data_et, output_shapes[0]);
 }
 
 std::shared_ptr<Node> BatchToSpace::clone_with_new_inputs(const OutputVector& new_args) const {
diff --git a/src/core/src/op/broadcast.cpp b/src/core/src/op/broadcast.cpp
index c7df0553d0b832..176e06aeec1e1f 100644
--- a/src/core/src/op/broadcast.cpp
+++ b/src/core/src/op/broadcast.cpp
@@ -48,8 +48,8 @@ std::pair<bool, ov::AxisSet> ov::op::v3::Broadcast::get_broadcast_axes() const {
         bool axes_known = false;
 
         if (get_input_partial_shape(0).is_static() && get_output_partial_shape(0).is_static()) {
-            const auto arg_shape = get_input_shape(0);
-            const auto result_shape = get_output_shape(0);
+            const auto& arg_shape = get_input_shape(0);
+            const auto& result_shape = get_output_shape(0);
             return get_broadcast_axes_bidirectional(arg_shape, result_shape);
         }
         return std::make_pair(axes_known, broadcast_axes);
@@ -108,7 +108,7 @@ ov::PartialShape get_result_shape_bidirectional(const ov::Node* this_ptr,
 
 bool ov::op::v3::Broadcast::broadcast_evaluate(ov::TensorVector& outputs, const ov::TensorVector& inputs) const {
     if (get_broadcast_spec().m_type == op::BroadcastType::BIDIRECTIONAL) {
-        auto arg_shape = inputs[0].get_shape();
+        const auto& arg_shape = inputs[0].get_shape();
         ov::Shape target_shape = op::util::BroadcastBase::get_target_shape(inputs[1]);
         ov::PartialShape result_shape =
             get_result_shape_bidirectional(this, ov::PartialShape{arg_shape}, ov::PartialShape{target_shape});
diff --git a/src/core/src/op/concat.cpp b/src/core/src/op/concat.cpp
index ccccb7c4f4bdcd..61c49d258b28a3 100644
--- a/src/core/src/op/concat.cpp
+++ b/src/core/src/op/concat.cpp
@@ -39,7 +39,8 @@ void Concat::validate_and_infer_types() {
         input_shapes.push_back(get_input_partial_shape(i));
     }
 
-    const auto output_shape = shape_infer(this, input_shapes).front();
+    const auto output_shapes = shape_infer(this, input_shapes);
+    const auto& output_shape = output_shapes[0];
     if (output_shape.rank().is_static() && (get_concatenation_axis() < 0)) {
         set_concatenation_axis(ov::util::normalize(get_axis(), output_shape.size()));
     }
diff --git a/src/core/src/op/ctc_loss.cpp b/src/core/src/op/ctc_loss.cpp
index a920e3b9c79c9f..6c82baa9a93490 100644
--- a/src/core/src/op/ctc_loss.cpp
+++ b/src/core/src/op/ctc_loss.cpp
@@ -59,8 +59,8 @@ void op::v4::CTCLoss::validate_and_infer_types() {
                               input_et);
     }
 
-    const auto output_shape = shape_infer(this, ov::util::get_node_input_partial_shapes(*this)).front();
-    set_output_type(0, logits_type, output_shape);
+    const auto output_shapes = shape_infer(this, ov::util::get_node_input_partial_shapes(*this));
+    set_output_type(0, logits_type, output_shapes[0]);
 }
 
 bool op::v4::CTCLoss::visit_attributes(AttributeVisitor& visitor) {
diff --git a/src/core/src/op/depth_to_space.cpp b/src/core/src/op/depth_to_space.cpp
index 0837897dac5604..b2f731d8666494 100644
--- a/src/core/src/op/depth_to_space.cpp
+++ b/src/core/src/op/depth_to_space.cpp
@@ -42,8 +42,8 @@ std::shared_ptr<Node> DepthToSpace::clone_with_new_inputs(const OutputVector& ne
 void DepthToSpace::validate_and_infer_types() {
     OV_OP_SCOPE(v0_DepthToSpace_validate_and_infer_types);
 
-    const auto output_shape = shape_infer(this, ov::util::get_node_input_partial_shapes(*this)).front();
-    set_output_type(0, get_input_element_type(0), output_shape);
+    const auto output_shapes = shape_infer(this, ov::util::get_node_input_partial_shapes(*this));
+    set_output_type(0, get_input_element_type(0), output_shapes[0]);
 }
 
 bool DepthToSpace::evaluate(TensorVector& outputs, const TensorVector& inputs) const {
diff --git a/src/core/src/op/gather_tree.cpp b/src/core/src/op/gather_tree.cpp
index a1618b598348da..f668fafbd8c72e 100644
--- a/src/core/src/op/gather_tree.cpp
+++ b/src/core/src/op/gather_tree.cpp
@@ -50,7 +50,7 @@ void op::v1::GatherTree::validate_and_infer_types() {
                           "Element type of inputs must be numeric. Got: ",
                           result_et);
 
-    const auto output_shape = shape_infer(this, ov::util::get_node_input_partial_shapes(*this)).front();
-    set_output_type(0, result_et, output_shape);
+    const auto output_shapes = shape_infer(this, ov::util::get_node_input_partial_shapes(*this));
+    set_output_type(0, result_et, output_shapes[0]);
 }
 }  // namespace ov
diff --git a/src/core/src/op/loop.cpp b/src/core/src/op/loop.cpp
index 3ad4e940012481..2d2ed726cea339 100644
--- a/src/core/src/op/loop.cpp
+++ b/src/core/src/op/loop.cpp
@@ -169,8 +169,8 @@ void Loop::validate_and_infer_types() {
 
             auto body_parameter = m_bodies[0]->get_parameters().at(merged_input_description->m_body_parameter_index);
 
-            auto input_partial_shape = input(index).get_partial_shape();
-            auto input_type = input(index).get_element_type();
+            const auto& input_partial_shape = input(index).get_partial_shape();
+            const auto& input_type = input(index).get_element_type();
 
             body_parameter->set_partial_shape(input_partial_shape);
             body_parameter->set_element_type(input_type);
diff --git a/src/core/src/op/max_pool.cpp b/src/core/src/op/max_pool.cpp
index 5ce63acfd9ac50..78df2bf78b8664 100644
--- a/src/core/src/op/max_pool.cpp
+++ b/src/core/src/op/max_pool.cpp
@@ -421,9 +421,9 @@ bool MaxPool::evaluate(TensorVector& outputs, const TensorVector& inputs) const
     const auto input_shapes = std::vector<PartialShape>{inputs[0].get_shape()};
     auto pads_begin = m_pads_begin;
     auto pads_end = m_pads_end;
-    const auto output_shape = shape_infer(this, input_shapes, pads_begin, pads_end).front();
+    const auto output_shapes = shape_infer(this, input_shapes, pads_begin, pads_end);
 
-    outputs[0].set_shape(output_shape.get_shape());
+    outputs[0].set_shape(output_shapes[0].get_shape());
 
     return ov::op::maxpool::evaluate_util(this, outputs, inputs, get_dilations(), get_axis());
 }
diff --git a/src/core/src/op/paged_attention.cpp b/src/core/src/op/paged_attention.cpp
index 909ed9b3ae7ea9..a9f5d8c0c6c459 100644
--- a/src/core/src/op/paged_attention.cpp
+++ b/src/core/src/op/paged_attention.cpp
@@ -14,15 +14,15 @@ PagedAttentionExtension::PagedAttentionExtension(const ov::OutputVector& args) :
 }
 
 void PagedAttentionExtension::validate_and_infer_types() {
-    auto value_cache_shape = get_input_partial_shape(4);
+    // const auto& value_cache_shape = get_input_partial_shape(4);
     // m_num_kv_heads = value_cache_shape[1];
     // m_head_size = value_cache_shape[2];
     // m_block_size = value_cache_shape[3];
     // Do not check shapes for cache K and cache V inputs, because they are hardware dependent
 
     // query: shape [batch_size, seq_len, num_heads * head_size]
-    auto query_type = get_input_element_type(0);
-    auto query_shape = get_input_partial_shape(0);
+    const auto& query_type = get_input_element_type(0);
+    const auto& query_shape = get_input_partial_shape(0);
     NODE_VALIDATION_CHECK(this,
                           // query_type.is_real() &&
                           query_shape.size() == 3,
@@ -34,8 +34,8 @@ void PagedAttentionExtension::validate_and_infer_types() {
                           query_shape);
 
     // key: shape [batch_size, seq_len, num_kv_heads * head_size]
-    auto key_type = get_input_element_type(1);
-    auto key_shape = get_input_partial_shape(1);
+    const auto& key_type = get_input_element_type(1);
+    const auto& key_shape = get_input_partial_shape(1);
     NODE_VALIDATION_CHECK(this,
                           // query_type == key_type &&
                           key_shape.size() == 3,
@@ -47,7 +47,6 @@ void PagedAttentionExtension::validate_and_infer_types() {
 
     // value: shape [batch_size, seq_len, num_kv_heads * head_size]
     // auto value_type = get_input_element_type(2);
-    auto value_shape = get_input_partial_shape(2);
 
     // is_prompt: boolean scalar
     NODE_VALIDATION_CHECK(this,
@@ -60,7 +59,7 @@ void PagedAttentionExtension::validate_and_infer_types() {
                           get_input_shape(5));
 
     // slot_mapping: shape [batch_size, max_context_len]
-    auto slot_mapping_shape = get_input_partial_shape(6);
+    const auto& slot_mapping_shape = get_input_partial_shape(6);
     NODE_VALIDATION_CHECK(this,
                           // get_input_element_type(6) == ov::element::i64 &&
                           slot_mapping_shape.size() == 2,
@@ -81,7 +80,7 @@ void PagedAttentionExtension::validate_and_infer_types() {
                           get_input_shape(7));
 
     // context_lens: shape [batch_size]
-    auto context_lens_shape = get_input_partial_shape(8);
+    const auto& context_lens_shape = get_input_partial_shape(8);
     NODE_VALIDATION_CHECK(this,
                           // get_input_element_type(8) == ov::element::i32 &&
                           context_lens_shape.size() == 1,
@@ -139,4 +138,4 @@ std::shared_ptr<ov::Node> PagedAttentionExtension::clone_with_new_inputs(const o
 }
 
 }  // namespace op
-}  // namespace ov
\ No newline at end of file
+}  // namespace ov
diff --git a/src/core/src/op/reverse.cpp b/src/core/src/op/reverse.cpp
index b778e204e2bb65..8763e2d116235e 100644
--- a/src/core/src/op/reverse.cpp
+++ b/src/core/src/op/reverse.cpp
@@ -63,8 +63,8 @@ void Reverse::validate_and_infer_types() {
                               "In 'index' mode the second input must contain integer values.");
     }
 
-    const auto output_shape = shape_infer(this, ov::util::get_node_input_partial_shapes(*this)).front();
-    set_output_type(0, get_input_element_type(0), output_shape);
+    const auto output_shapes = shape_infer(this, ov::util::get_node_input_partial_shapes(*this));
+    set_output_type(0, get_input_element_type(0), output_shapes[0]);
 }
 
 std::shared_ptr<ov::Node> Reverse::clone_with_new_inputs(const OutputVector& new_args) const {
diff --git a/src/core/src/op/reverse_sequence.cpp b/src/core/src/op/reverse_sequence.cpp
index 36d7dec3f151ba..2b584996d3fa7d 100644
--- a/src/core/src/op/reverse_sequence.cpp
+++ b/src/core/src/op/reverse_sequence.cpp
@@ -38,8 +38,8 @@ void op::v0::ReverseSequence::validate_and_infer_types() {
                           "Sequence lengths element type must be numeric type. Got: ",
                           seq_lengths_et);
 
-    const auto output_shape = shape_infer(this, ov::util::get_node_input_partial_shapes(*this)).front();
-    set_output_type(0, get_input_element_type(0), output_shape);
+    const auto output_shapes = shape_infer(this, ov::util::get_node_input_partial_shapes(*this));
+    set_output_type(0, get_input_element_type(0), output_shapes[0]);
 
     m_normalized_seq_axis = ov::util::normalize_axis(this, m_seq_axis, get_input_partial_shape(0).rank());
 }
diff --git a/src/core/src/op/roll.cpp b/src/core/src/op/roll.cpp
index d4ec7fdb0ecaa1..4ce4e57d527c52 100644
--- a/src/core/src/op/roll.cpp
+++ b/src/core/src/op/roll.cpp
@@ -28,8 +28,8 @@ void Roll::validate_and_infer_types() {
                           axes_et.is_dynamic() || axes_et == element::i32 || axes_et == element::i64,
                           "Axes must have int32 or int64 element type.");
 
-    const auto output_shape = shape_infer(this, ov::util::get_node_input_partial_shapes(*this)).front();
-    set_output_type(0, get_input_element_type(0), output_shape);
+    const auto output_shapes = shape_infer(this, ov::util::get_node_input_partial_shapes(*this));
+    set_output_type(0, get_input_element_type(0), output_shapes[0]);
 }
 
 bool Roll::visit_attributes(AttributeVisitor& visitor) {
diff --git a/src/core/src/op/scaled_dot_product_attention.cpp b/src/core/src/op/scaled_dot_product_attention.cpp
index 61bdde3324d3d5..cdce3bfbd5ff4b 100644
--- a/src/core/src/op/scaled_dot_product_attention.cpp
+++ b/src/core/src/op/scaled_dot_product_attention.cpp
@@ -64,7 +64,7 @@ void op::v13::ScaledDotProductAttention::validate_and_infer_types() {
                           "The element type of the input tensor must be a floating-point.");
 
     const auto& input_shapes = ov::util::get_node_input_partial_shapes(*this);
-    const auto& output_shapes = shape_infer(this, input_shapes);
+    const auto output_shapes = shape_infer(this, input_shapes);
     set_output_type(0, out_type, output_shapes[0]);
 }
 
diff --git a/src/core/src/op/shape_of.cpp b/src/core/src/op/shape_of.cpp
index 07138610a4cc12..1c12b54b1fdbe9 100644
--- a/src/core/src/op/shape_of.cpp
+++ b/src/core/src/op/shape_of.cpp
@@ -120,7 +120,7 @@ void ShapeOf::validate_and_infer_types() {
                           m_output_type == element::i64 || m_output_type == element::i32,
                           "Output type must be i32 or i64");
     set_input_is_relevant_to_value(0, false);
-    const auto input_partial_shape = get_input_partial_shape(0);
+    const auto& input_partial_shape = get_input_partial_shape(0);
     set_output_type(0, m_output_type, PartialShape{input_partial_shape.rank()});
 }
 
diff --git a/src/core/src/op/shuffle_channels.cpp b/src/core/src/op/shuffle_channels.cpp
index 80658a696e3585..3927d260309b15 100644
--- a/src/core/src/op/shuffle_channels.cpp
+++ b/src/core/src/op/shuffle_channels.cpp
@@ -41,8 +41,8 @@ size_t ShuffleChannels::get_zero_based_axis() const {
 void ShuffleChannels::validate_and_infer_types() {
     OV_OP_SCOPE(v0_ShuffleChannels_validate_and_infer_types);
 
-    const auto output_shape = shape_infer(this, ov::util::get_node_input_partial_shapes(*this)).front();
-    set_output_type(0, get_input_element_type(0), output_shape);
+    const auto output_shapes = shape_infer(this, ov::util::get_node_input_partial_shapes(*this));
+    set_output_type(0, get_input_element_type(0), output_shapes[0]);
 }
 
 std::shared_ptr<Node> ShuffleChannels::clone_with_new_inputs(const OutputVector& new_args) const {
diff --git a/src/core/src/op/slice.cpp b/src/core/src/op/slice.cpp
index a1143da2afa5ae..519663a083ccc8 100644
--- a/src/core/src/op/slice.cpp
+++ b/src/core/src/op/slice.cpp
@@ -125,7 +125,7 @@ bool Slice::evaluate(TensorVector& outputs, const TensorVector& inputs) const {
 
     const auto starts = ov::get_tensor_data_as<int64_t>(inputs[1]);
     const auto steps = ov::get_tensor_data_as<int64_t>(inputs[3]);
-    const auto axes = slice_no_axes(this) ? default_axes(starts.size()) : ov::get_tensor_data_as<int64_t>(inputs[4]);
+    const auto& axes = slice_no_axes(this) ? default_axes(starts.size()) : ov::get_tensor_data_as<int64_t>(inputs[4]);
 
     reference::slice(static_cast<const char*>(inputs[0].data()),
                      inputs[0].get_shape(),
diff --git a/src/core/src/op/space_to_batch.cpp b/src/core/src/op/space_to_batch.cpp
index 7d7237a4e0a25a..a641586c6879aa 100644
--- a/src/core/src/op/space_to_batch.cpp
+++ b/src/core/src/op/space_to_batch.cpp
@@ -55,8 +55,8 @@ void SpaceToBatch::validate_and_infer_types() {
                           "pads_end must be an integral number but got (",
                           pads_end_type,
                           ").");
-    const auto output_shape = shape_infer(this, ov::util::get_node_input_partial_shapes(*this)).front();
-    set_output_type(0, data_type, output_shape);
+    const auto output_shapes = shape_infer(this, ov::util::get_node_input_partial_shapes(*this));
+    set_output_type(0, data_type, output_shapes[0]);
 }
 
 std::shared_ptr<Node> SpaceToBatch::clone_with_new_inputs(const OutputVector& new_args) const {
diff --git a/src/core/src/op/space_to_depth.cpp b/src/core/src/op/space_to_depth.cpp
index 43ef62906a2dee..e1a5b8997127fd 100644
--- a/src/core/src/op/space_to_depth.cpp
+++ b/src/core/src/op/space_to_depth.cpp
@@ -46,8 +46,8 @@ std::shared_ptr<Node> SpaceToDepth::clone_with_new_inputs(const OutputVector& ne
 void SpaceToDepth::validate_and_infer_types() {
     OV_OP_SCOPE(v0_SpaceToDepth_validate_and_infer_types);
 
-    const auto output_shape = shape_infer(this, ov::util::get_node_input_partial_shapes(*this)).front();
-    set_output_type(0, get_input_element_type(0), output_shape);
+    const auto output_shapes = shape_infer(this, ov::util::get_node_input_partial_shapes(*this));
+    set_output_type(0, get_input_element_type(0), output_shapes[0]);
 }
 
 bool SpaceToDepth::evaluate(TensorVector& outputs, const TensorVector& inputs) const {
diff --git a/src/core/src/op/strided_slice.cpp b/src/core/src/op/strided_slice.cpp
index 62049ba8938f8a..deb89fa9a531d4 100644
--- a/src/core/src/op/strided_slice.cpp
+++ b/src/core/src/op/strided_slice.cpp
@@ -48,8 +48,8 @@ StridedSlice::StridedSlice(const Output<Node>& data,
 
 namespace {
 std::shared_ptr<Node> calculate_default_strides(const Output<Node>& begin, const Output<Node>& end) {
-    const auto begin_pshape = begin.get_partial_shape();
-    const auto end_pshape = end.get_partial_shape();
+    const auto& begin_pshape = begin.get_partial_shape();
+    const auto& end_pshape = end.get_partial_shape();
 
     size_t strides_length = 0;
     if (begin_pshape.rank().is_static() && begin_pshape.rank().get_length() == 1 && begin_pshape[0].is_static()) {
diff --git a/src/core/src/op/tensor_iterator.cpp b/src/core/src/op/tensor_iterator.cpp
index 255b49b536070b..0c98ffd61d3e72 100644
--- a/src/core/src/op/tensor_iterator.cpp
+++ b/src/core/src/op/tensor_iterator.cpp
@@ -106,14 +106,12 @@ void op::v0::TensorIterator::validate_and_infer_types() {
 
             auto body_parameter = m_bodies[0]->get_parameters().at(merged_input_description->m_body_parameter_index);
 
-            auto body_param_partial_shape = body_parameter->get_partial_shape();
-            auto input_partial_shape = inputs().at(index).get_source_output().get_partial_shape();
+            const auto& input_partial_shape = inputs().at(index).get_source_output().get_partial_shape();
             body_parameter->set_partial_shape(input_partial_shape);
         } else if (auto invariant_input_description = ov::as_type_ptr<InvariantInputDescription>(input_description)) {
             auto body_parameter = m_bodies[0]->get_parameters().at(invariant_input_description->m_body_parameter_index);
 
-            auto body_param_partial_shape = body_parameter->get_partial_shape();
-            auto input_partial_shape = inputs().at(index).get_source_output().get_partial_shape();
+            const auto& input_partial_shape = inputs().at(index).get_source_output().get_partial_shape();
             body_parameter->set_partial_shape(input_partial_shape);
         }
     }
diff --git a/src/core/src/op/unique.cpp b/src/core/src/op/unique.cpp
index e11a73381a347a..8088f1fa7098d4 100644
--- a/src/core/src/op/unique.cpp
+++ b/src/core/src/op/unique.cpp
@@ -142,7 +142,7 @@ void op::v10::Unique::validate_and_infer_types() {
 
             if (input_shape.rank().is_static()) {
                 const auto normalized_axis = ov::util::normalize_axis(this, axis, input_shape.rank());
-                const auto dim_at_axis = input_shape[normalized_axis];
+                const auto& dim_at_axis = input_shape[normalized_axis];
 
                 Dimension output_dim_at_axis;
                 Dimension rev_idx_size;
diff --git a/src/core/src/op/util/broadcast_base.cpp b/src/core/src/op/util/broadcast_base.cpp
index 48ca06dbd97f06..1d4f7c1541ae1f 100644
--- a/src/core/src/op/util/broadcast_base.cpp
+++ b/src/core/src/op/util/broadcast_base.cpp
@@ -218,8 +218,8 @@ void ov::op::util::BroadcastBase::validate_and_infer_types() {
         // Validate axes_mapping
         if (get_input_partial_shape(0).is_static() && get_input_partial_shape(1).is_static() &&
             get_input_partial_shape(2).is_static()) {
-            auto arg_shape = get_input_shape(0);
-            auto axes_shape = get_input_shape(2);
+            const auto& arg_shape = get_input_shape(0);
+            const auto& axes_shape = get_input_shape(2);
             auto input_rank = (arg_shape.size() == 0 && shape_size(axes_shape) > 0) ? 1 : arg_shape.size();
 
             // Rank(arg_shape) == shape_size(axes_mapping)
@@ -297,8 +297,8 @@ std::pair<bool, ov::AxisSet> ov::op::util::BroadcastBase::get_broadcast_axes() c
         }
     } else if (m_mode.m_type == BroadcastType::NUMPY || m_mode.m_type == BroadcastType::PDPD) {
         if (get_input_partial_shape(0).is_static() && get_output_partial_shape(0).is_static()) {
-            auto arg_shape = get_input_shape(0);
-            auto result_shape = get_output_shape(0);
+            const auto& arg_shape = get_input_shape(0);
+            const auto& result_shape = get_output_shape(0);
             return get_broadcast_axes_numpy_pdpd(arg_shape, result_shape, m_mode);
         }
     } else {
@@ -435,7 +435,7 @@ bool ov::op::util::BroadcastBase::evaluate(ov::TensorVector& outputs, const ov::
 
     PartialShape result_shape;
     std::pair<bool, AxisSet> pair_broadcast_axes;
-    auto arg_shape = inputs[0].get_shape();
+    const auto& arg_shape = inputs[0].get_shape();
 
     if (m_mode.m_type == BroadcastType::NONE) {
         AxisVector axes_mapping_val;
diff --git a/src/core/src/op/util/gather_base.cpp b/src/core/src/op/util/gather_base.cpp
index 8e7b9fe4531343..92e41781b1de55 100644
--- a/src/core/src/op/util/gather_base.cpp
+++ b/src/core/src/op/util/gather_base.cpp
@@ -49,7 +49,7 @@ bool cf_gather_with_subgraph(OutputVector& output_values,
         return false;
     }
     // only single indices are accepted
-    const auto indices_shape = indices->get_shape();
+    const auto& indices_shape = indices->get_shape();
     if (indices_shape.size() > 1 || (indices_shape.size() == 1 && indices_shape[0] > 1)) {
         return false;
     }
diff --git a/src/core/src/op/util/multi_subgraph_base.cpp b/src/core/src/op/util/multi_subgraph_base.cpp
index 43cd61e6645b2b..144cf167057543 100644
--- a/src/core/src/op/util/multi_subgraph_base.cpp
+++ b/src/core/src/op/util/multi_subgraph_base.cpp
@@ -152,7 +152,7 @@ void ov::op::util::MultiSubGraphOp::validate_and_infer_type_body(
         auto index = input_description->m_input_index;
 
         auto body_parameter = params.at(input_description->m_body_parameter_index);
-        auto input_partial_shape = input_value(index).get_partial_shape();
+        const auto& input_partial_shape = input_value(index).get_partial_shape();
         auto dtype = input_value(index).get_element_type();
         body_parameter->set_partial_shape(input_partial_shape);
         body_parameter->set_element_type(dtype);
diff --git a/src/core/src/op/util/scatter_elements_update_base.cpp b/src/core/src/op/util/scatter_elements_update_base.cpp
index 9126e5cda930b4..75be89e6c5bd99 100644
--- a/src/core/src/op/util/scatter_elements_update_base.cpp
+++ b/src/core/src/op/util/scatter_elements_update_base.cpp
@@ -44,11 +44,11 @@ void util::ScatterElementsUpdateBase::validate_and_infer_types() {
                           data_et,
                           " and: ",
                           updates_et);
-    const auto output_shape = shape_infer(this, ov::util::get_node_input_partial_shapes(*this)).front();
+    const auto output_shapes = shape_infer(this, ov::util::get_node_input_partial_shapes(*this));
     auto out_et = get_input_element_type(0);
     std::ignore = element::Type::merge(out_et, get_input_element_type(0), get_input_element_type(2));
-    set_output_type(0, out_et, output_shape);
-    if (output_shape.is_dynamic()) {
+    set_output_type(0, out_et, output_shapes[0]);
+    if (output_shapes[0].is_dynamic()) {
         set_input_is_relevant_to_shape(0);
     }
 }
diff --git a/src/core/src/partial_shape.cpp b/src/core/src/partial_shape.cpp
index 6f503aa608833f..af346407e32636 100644
--- a/src/core/src/partial_shape.cpp
+++ b/src/core/src/partial_shape.cpp
@@ -314,8 +314,8 @@ bool ov::PartialShape::broadcast_merge_into(PartialShape& dst,
             std::vector<Dimension> dims(new_rank);
             bool success = true;
             for (int64_t i = 0; i < new_rank; i++) {
-                auto dsti = i < (new_rank - dst_rank) ? Dimension(1) : dst[i - (new_rank - dst_rank)];
-                auto srci = i < (new_rank - src_rank) ? Dimension(1) : src[i - (new_rank - src_rank)];
+                const auto& dsti = i < (new_rank - dst_rank) ? Dimension(1) : dst[i - (new_rank - dst_rank)];
+                const auto& srci = i < (new_rank - src_rank) ? Dimension(1) : src[i - (new_rank - src_rank)];
                 success &= Dimension::broadcast_merge(dims[i], dsti, srci);
             }
             dst = PartialShape(std::move(dims));
diff --git a/src/core/src/pass/constant_folding.cpp b/src/core/src/pass/constant_folding.cpp
index c691c818a9ce7b..3de91829f91b0c 100644
--- a/src/core/src/pass/constant_folding.cpp
+++ b/src/core/src/pass/constant_folding.cpp
@@ -39,14 +39,13 @@ const auto is_output_foldable = [](const ov::Output<ov::Node>& output) {
  *
  * \return std::string with new friendly name.
  */
-const auto friendly_name_from = [](const ov::Node& node, const size_t output_count, const size_t idx) {
+const auto friendly_name_from = [](const ov::Node& node, const size_t output_count, const size_t idx) -> std::string {
     constexpr auto single_output = static_cast<size_t>(1);
-
-    if (single_output == output_count) {
-        return node.get_friendly_name();
-    } else {
-        return node.get_friendly_name() + "." + std::to_string(idx);
+    auto name = node.get_friendly_name();
+    if (single_output != output_count) {
+        name.append(".").append(std::to_string(idx));
     }
+    return name;
 };
 
 static bool restore_original_input_precision(const std::shared_ptr<ov::Node>& node) {
@@ -128,7 +127,7 @@ bool ov::pass::ConstantFolding::run_on_model(const std::shared_ptr<ov::Model>& m
 
             for (size_t i = 0; i < replacements.size(); ++i) {
                 auto node_output = original_node->output(i);
-                auto replacement = replacements.at(i);
+                const auto& replacement = replacements.at(i);
                 auto replacement_ptr = replacement.get_node_shared_ptr();
                 if (replacement_ptr && (node_output != replacement)) {
                     replacement_ptr->set_friendly_name(friendly_name_from(*original_node, replacements.size(), i));
diff --git a/src/core/src/pass/low_latency.cpp b/src/core/src/pass/low_latency.cpp
index 4590f8c876aaef..2dd75aa072e4fb 100644
--- a/src/core/src/pass/low_latency.cpp
+++ b/src/core/src/pass/low_latency.cpp
@@ -128,7 +128,7 @@ std::vector<std::shared_ptr<ov::opset9::Assign>> replace_with_memory(const std::
 }
 
 bool need_unroll(const std::shared_ptr<ov::Node>& op) {
-    const auto p_shape = op->get_input_partial_shape(0);
+    const auto& p_shape = op->get_input_partial_shape(0);
     if (p_shape.rank().is_dynamic() || p_shape[1].is_dynamic() || p_shape[1].get_length() != 1) {
         return false;
     }
diff --git a/src/core/src/pass/serialize.cpp b/src/core/src/pass/serialize.cpp
index c82c5a806578f9..f38500bc3541e8 100644
--- a/src/core/src/pass/serialize.cpp
+++ b/src/core/src/pass/serialize.cpp
@@ -1033,7 +1033,7 @@ void ngfunction_2_ir(pugi::xml_node& netXml,
                 pugi::xml_node port = output.append_child("port");
                 port.append_attribute("id").set_value(port_id++);
 
-                auto rt_info = o.get_tensor().get_rt_info();
+                const auto& rt_info = o.get_tensor().get_rt_info();
                 auto port_element_type =
                     is_fp16_compression_postponed(rt_info) ? ov::element::f16 : o.get_element_type();
 
@@ -1054,7 +1054,7 @@ void ngfunction_2_ir(pugi::xml_node& netXml,
                     port.append_attribute("names").set_value(names.c_str());
                 }
 
-                for (auto d : o.get_partial_shape()) {
+                for (const auto& d : o.get_partial_shape()) {
                     pugi::xml_node dim = port.append_child("dim");
                     if (d.is_dynamic()) {
                         dim.append_child(pugi::xml_node_type::node_pcdata).set_value("-1");
diff --git a/src/core/src/pass/visualize_tree.cpp b/src/core/src/pass/visualize_tree.cpp
index c8656e368cf5e5..61446c416d132d 100644
--- a/src/core/src/pass/visualize_tree.cpp
+++ b/src/core/src/pass/visualize_tree.cpp
@@ -611,7 +611,7 @@ std::string ov::pass::VisualizeTree::get_node_name(std::shared_ptr<Node> node) {
     if (node->get_friendly_name() != node->get_name()) {
         rc += "\\n" + (nvtmn ? std::string("name: ") : "") + node->get_name();
     }
-    const auto type_info = node->get_type_info();
+    const auto& type_info = node->get_type_info();
     rc += "\\n" + (nvtmn ? std::string("type_name: ") : "") + std::string(type_info.version_id) +
           "::" + std::string(type_info.name);
 
@@ -655,7 +655,7 @@ std::string ov::pass::VisualizeTree::get_node_name(std::shared_ptr<Node> node) {
 
     static const bool nvtrti = ov::util::getenv_bool("OV_VISUALIZE_TREE_RUNTIME_INFO");
     if (nvtrti) {
-        const auto rt = node->get_rt_info();
+        const auto& rt = node->get_rt_info();
         if (!rt.empty()) {
             rc += "\\nrt info: " + get_attribute_values(rt, "\\n");
         }
diff --git a/src/core/src/preprocess/preprocess_impls.cpp b/src/core/src/preprocess/preprocess_impls.cpp
index bcc511332a06bc..a3a3c36291fc22 100644
--- a/src/core/src/preprocess/preprocess_impls.cpp
+++ b/src/core/src/preprocess/preprocess_impls.cpp
@@ -44,7 +44,7 @@ InputInfo::InputInfoImpl::InputInfoData InputInfo::InputInfoImpl::create_new_par
         }
     }
 
-    auto model_shape = res.m_param->get_partial_shape();
+    const auto& model_shape = res.m_param->get_partial_shape();
     auto new_param_shape = model_shape;
     if (get_tensor_data()->is_shape_set()) {
         new_param_shape = get_tensor_data()->get_shape();
@@ -63,9 +63,12 @@ InputInfo::InputInfoImpl::InputInfoData InputInfo::InputInfoImpl::create_new_par
         auto net_to_tensor = layout::utils::find_permutation(sq_layout, new_param_shape, res.m_tensor_layout);
         if (!net_to_tensor.empty() && new_param_shape.rank().is_static()) {
             std::vector<ov::Dimension> dims(new_param_shape.size());
-            std::transform(net_to_tensor.begin(), net_to_tensor.end(), dims.begin(), [&](int64_t v) {
-                return new_param_shape[v];
-            });
+            std::transform(net_to_tensor.begin(),
+                           net_to_tensor.end(),
+                           dims.begin(),
+                           [&](int64_t v) -> const Dimension& {
+                               return new_param_shape[v];
+                           });
             new_param_shape = PartialShape(dims);
         }
     } else {
@@ -191,7 +194,7 @@ bool InputInfo::InputInfoImpl::build(const std::shared_ptr<Model>& model,
         nodes = std::get<0>(action_result);
     }
 
-    auto node = nodes[0];
+    const auto& node = nodes[0];
     if (node.get_partial_shape() != context.model_shape()) {
         need_validate = true;  // Trigger revalidation if input parameter shape is changed
     }
@@ -324,7 +327,7 @@ void InputInfo::InputInfoImpl::dump(std::ostream& str,
 void OutputInfo::OutputInfoImpl::build(ov::ResultVector& results) {
     std::shared_ptr<opset8::Result> result;
     auto node = m_output_node;
-    auto start_out_node_names = node.get_tensor().get_names();
+    const auto start_out_node_names = node.get_tensor().get_names();
     node.get_tensor().set_names({});
     result = std::dynamic_pointer_cast<opset8::Result>(node.get_node_shared_ptr());
     // Set result layout from 'model' information
@@ -422,7 +425,7 @@ void OutputInfo::OutputInfoImpl::build(ov::ResultVector& results) {
 void OutputInfo::OutputInfoImpl::dump(std::ostream& str) const {
     std::shared_ptr<opset8::Result> result;
     auto node = m_output_node;
-    auto start_out_node_names = node.get_tensor().get_names();
+    const auto& start_out_node_names = node.get_tensor().get_names();
     result = std::dynamic_pointer_cast<opset8::Result>(node.get_node_shared_ptr());
     auto model_layout = get_model_data()->is_layout_set() ? get_model_data()->get_layout() : result->get_layout();
     PostprocessingContext context(model_layout);
diff --git a/src/core/src/runtime/string_aligned_buffer.cpp b/src/core/src/runtime/string_aligned_buffer.cpp
index 98503e0fb507f6..cc2bcdd5b0eac7 100644
--- a/src/core/src/runtime/string_aligned_buffer.cpp
+++ b/src/core/src/runtime/string_aligned_buffer.cpp
@@ -58,7 +58,7 @@ void aux_get_header(const std::shared_ptr<ov::StringAlignedBuffer>& string_align
     size_t current_symbols_pos = 0;
 
     for (size_t ind = 0; ind < num_elements; ++ind) {
-        auto str = strings[ind];
+        const auto& str = strings[ind];
         current_symbols_pos += str.size();
         *pindices = int32_t(current_symbols_pos);
         ++pindices;
diff --git a/src/core/template_extension/identity.cpp b/src/core/template_extension/identity.cpp
index f3cf38ba1b2141..4b6f6f12bafcde 100644
--- a/src/core/template_extension/identity.cpp
+++ b/src/core/template_extension/identity.cpp
@@ -35,8 +35,8 @@ bool Identity::visit_attributes(ov::AttributeVisitor& visitor) {
 
 //! [op:evaluate]
 bool Identity::evaluate(ov::TensorVector& outputs, const ov::TensorVector& inputs) const {
-    auto in = inputs[0];
-    auto out = outputs[0];
+    const auto& in = inputs[0];
+    auto& out = outputs[0];
     if (out.data() == in.data())  // Nothing to do
         return true;
     out.set_shape(in.get_shape());
diff --git a/src/core/tests/visitors/visitors.hpp b/src/core/tests/visitors/visitors.hpp
index 2d78c466b9a146..d4c3c54994ef60 100644
--- a/src/core/tests/visitors/visitors.hpp
+++ b/src/core/tests/visitors/visitors.hpp
@@ -489,7 +489,6 @@ class NodeBuilder : public ValueMap, public DeserializeAttributeVisitor {
 }  // namespace test
 }  // namespace ov
 
-
 #ifdef __GNUC__
 #    pragma GCC diagnostic pop
 #endif