diff --git a/src/plugins/intel_gpu/src/graph/impls/common/loop.cpp b/src/plugins/intel_gpu/src/graph/impls/common/loop.cpp
index ada8d22bbcb731..2d7b6417e3907b 100644
--- a/src/plugins/intel_gpu/src/graph/impls/common/loop.cpp
+++ b/src/plugins/intel_gpu/src/graph/impls/common/loop.cpp
@@ -2,6 +2,7 @@
 // SPDX-License-Identifier: Apache-2.0
 //
 #include "intel_gpu/graph/kernel_impl_params.hpp"
+#include "loop.hpp"
 #include "loop_inst.h"
 #include "registry/implementation_map.hpp"
 #include "register.hpp"
@@ -304,6 +305,11 @@ struct loop_impl : typed_primitive_impl<loop> {
     std::vector<cldnn::loop::backedge_mapping> _back_edges;
 };
 
+std::unique_ptr<primitive_impl> LoopImplementationManager::create_impl(const program_node& node, const kernel_impl_params& params) const {
+    assert(node.is_type<loop>());
+    return loop_impl::create(static_cast<const loop_node&>(node), params);
+}
+
 namespace detail {
 attach_loop_common::attach_loop_common() {
     implementation_map<loop>::add(impl_types::common,
diff --git a/src/plugins/intel_gpu/src/graph/impls/common/loop.hpp b/src/plugins/intel_gpu/src/graph/impls/common/loop.hpp
new file mode 100644
index 00000000000000..be94548705972f
--- /dev/null
+++ b/src/plugins/intel_gpu/src/graph/impls/common/loop.hpp
@@ -0,0 +1,36 @@
+// Copyright (C) 2024 Intel Corporation
+// SPDX-License-Identifier: Apache-2.0
+//
+
+#include "registry/implementation_manager.hpp"
+#include "program_node.h"
+
+#include <memory>
+
+namespace cldnn {
+
+namespace common {
+
+struct LoopImplementationManager : public ImplementationManager {
+    OV_GPU_PRIMITIVE_IMPL("common::loop")
+    LoopImplementationManager(shape_types shape_type, ValidateFunc vf = nullptr) : ImplementationManager(impl_types::common, shape_type, vf) {}
+
+    std::unique_ptr<primitive_impl> create_impl(const program_node& node, const kernel_impl_params& params) const override;
+
+    in_out_fmts_t query_formats(const program_node& node) const override {
+        std::vector<format::type> in_fmts(node.get_dependencies().size(), format::any);
+        std::vector<format::type> out_fmts(node.get_outputs_count(), format::any);
+
+        for (size_t i = 0; i < node.get_dependencies().size(); i++) {
+            size_t in_rank = node.get_input_layout(i).get_rank();
+            in_fmts[i] = format::get_default_format(in_rank);
+        }
+        size_t out_rank = node.get_output_layout().get_rank();
+        out_fmts[0] = format::get_default_format(out_rank);
+
+        return {in_fmts, out_fmts};
+    }
+};
+
+} // namespace common
+} // namespace cldnn
diff --git a/src/plugins/intel_gpu/src/graph/registry/loop_impls.cpp b/src/plugins/intel_gpu/src/graph/registry/loop_impls.cpp
new file mode 100644
index 00000000000000..ec92b335c55928
--- /dev/null
+++ b/src/plugins/intel_gpu/src/graph/registry/loop_impls.cpp
@@ -0,0 +1,27 @@
+// Copyright (C) 2024 Intel Corporation
+// SPDX-License-Identifier: Apache-2.0
+//
+
+#include "registry.hpp"
+#include "intel_gpu/primitives/loop.hpp"
+#include "primitive_inst.h"
+
+#if OV_GPU_WITH_COMMON
+    #include "impls/common/loop.hpp"
+#endif
+
+
+namespace ov::intel_gpu {
+
+using namespace cldnn;
+
+const std::vector<std::shared_ptr<cldnn::ImplementationManager>>& Registry<loop>::get_implementations() {
+    static const std::vector<std::shared_ptr<ImplementationManager>> impls = {
+        OV_GPU_CREATE_INSTANCE_COMMON(common::LoopImplementationManager, shape_types::static_shape)
+        OV_GPU_CREATE_INSTANCE_COMMON(common::LoopImplementationManager, shape_types::dynamic_shape)
+    };
+
+    return impls;
+}
+
+}  // namespace ov::intel_gpu
diff --git a/src/plugins/intel_gpu/src/graph/registry/registry.hpp b/src/plugins/intel_gpu/src/graph/registry/registry.hpp
index 1a57a1ed1ac970..44f070f95cc60b 100644
--- a/src/plugins/intel_gpu/src/graph/registry/registry.hpp
+++ b/src/plugins/intel_gpu/src/graph/registry/registry.hpp
@@ -79,8 +79,10 @@
 #endif
 
 #if OV_GPU_WITH_COMMON
-#    define OV_GPU_GET_INSTANCE_COMMON(prim, ...) EXPAND(GET_INSTANCE(prim, cldnn::impl_types::common, __VA_ARGS__))
+#    define OV_GPU_CREATE_INSTANCE_COMMON(...) EXPAND(CREATE_INSTANCE(__VA_ARGS__))
+#    define OV_GPU_GET_INSTANCE_COMMON(prim, ...) EXPAND(SELECT(COUNT(__VA_ARGS__), prim, impl_types::ocl, __VA_ARGS__))
 #else
+#    define OV_GPU_CREATE_INSTANCE_COMMON(...)
 #    define OV_GPU_GET_INSTANCE_COMMON(...)
 #endif
 
@@ -138,6 +140,7 @@ REGISTER_IMPLS(gather);
 REGISTER_IMPLS(gather_nd);
 REGISTER_IMPLS(gemm);
 REGISTER_IMPLS(group_normalization);
+REGISTER_IMPLS(loop);
 REGISTER_IMPLS(lora);
 REGISTER_IMPLS(lstm_cell);
 REGISTER_IMPLS(lstm_seq);
@@ -162,7 +165,6 @@ REGISTER_IMPLS(col2im);
 REGISTER_DEFAULT_IMPLS(assign, CPU_S, CPU_D);
 REGISTER_DEFAULT_IMPLS(read_value, CPU_S, CPU_D);
 REGISTER_DEFAULT_IMPLS(condition, COMMON_S, COMMON_D);
-REGISTER_DEFAULT_IMPLS(loop, COMMON_S, COMMON_D);
 REGISTER_DEFAULT_IMPLS(input_layout, COMMON_S, COMMON_D);
 REGISTER_DEFAULT_IMPLS(non_max_suppression_gather, CPU_S);
 REGISTER_DEFAULT_IMPLS(proposal, CPU_S, CPU_D);
diff --git a/src/plugins/intel_gpu/tests/unit/module_tests/impls_registry_test.cpp b/src/plugins/intel_gpu/tests/unit/module_tests/impls_registry_test.cpp
index 9e9655ce4e3c83..28b563a7509c99 100644
--- a/src/plugins/intel_gpu/tests/unit/module_tests/impls_registry_test.cpp
+++ b/src/plugins/intel_gpu/tests/unit/module_tests/impls_registry_test.cpp
@@ -43,7 +43,6 @@
 #include "intel_gpu/primitives/grn.hpp"
 #include "intel_gpu/primitives/group_normalization.hpp"
 #include "intel_gpu/primitives/kv_cache.hpp"
-#include "intel_gpu/primitives/loop.hpp"
 #include "intel_gpu/primitives/matrix_nms.hpp"
 #include "intel_gpu/primitives/multiclass_nms.hpp"
 #include "intel_gpu/primitives/multinomial.hpp"
@@ -145,7 +144,6 @@ TEST(registry_test, no_null_impls) {
                 cldnn::assign,
                 cldnn::read_value,
                 cldnn::condition,
-                cldnn::loop,
                 cldnn::input_layout,
                 cldnn::non_max_suppression_gather,
                 cldnn::proposal,