diff --git a/tests/validation/CL/Winograd.cpp b/tests/validation/CL/Winograd.cpp
index 8fa5826..08810ea 100644
--- a/tests/validation/CL/Winograd.cpp
+++ b/tests/validation/CL/Winograd.cpp
@@ -201,19 +201,23 @@
     winograd_filter_transform.configure(&a, &b, winograd_info);
 }
 
-FIXTURE_DATA_TEST_CASE(RunSmall, CLWinogradFilterTransformFixture, framework::DatasetMode::ALL, combine(combine(combine(datasets::Small3x3Shapes(),
-                                                                                                                        framework::dataset::make("OutputTile", { Size2D(2U, 2U), Size2D(4U, 4U) })),
-                                                                                                                framework::dataset::make("DataLayout", { DataLayout::NCHW })),
-                                                                                                        framework::dataset::make("DataType", { DataType::F32 })))
+FIXTURE_DATA_TEST_CASE(RunSmall, CLWinogradFilterTransformFixture, framework::DatasetMode::ALL,
+                       combine(combine(framework::dataset::concat(framework::dataset::concat(combine(datasets::Small3x3Shapes(), framework::dataset::make("OutputTile", Size2D(2U, 2U))), combine(datasets::Small3x3Shapes(),
+                                                                                             framework::dataset::make("OutputTile", Size2D(4U, 4U)))),
+                                                                  combine(datasets::Small5x5Shapes(), framework::dataset::make("OutputTile", Size2D(4U, 4U)))),
+                                       framework::dataset::make("DataLayout", { DataLayout::NCHW })),
+                               framework::dataset::make("DataType", { DataType::F32 })))
 {
     // Validate output
     validate(CLAccessor(_target), _reference, tolerance_f32);
 }
 
-FIXTURE_DATA_TEST_CASE(RunLarge, CLWinogradFilterTransformFixture, framework::DatasetMode::NIGHTLY, combine(combine(combine(datasets::Large3x3Shapes(),
-                                                                                                                    framework::dataset::make("OutputTile", { Size2D(2U, 2U), Size2D(4U, 4U) })),
-                                                                                                                    framework::dataset::make("DataLayout", { DataLayout::NCHW })),
-                                                                                                            framework::dataset::make("DataType", { DataType::F32 })))
+FIXTURE_DATA_TEST_CASE(RunLarge, CLWinogradFilterTransformFixture, framework::DatasetMode::NIGHTLY,
+                       combine(combine(framework::dataset::concat(framework::dataset::concat(combine(datasets::Large3x3Shapes(), framework::dataset::make("OutputTile", Size2D(2U, 2U))), combine(datasets::Large3x3Shapes(),
+                                                                                             framework::dataset::make("OutputTile", Size2D(4U, 4U)))),
+                                                                  combine(datasets::Large5x5Shapes(), framework::dataset::make("OutputTile", Size2D(4U, 4U)))),
+                                       framework::dataset::make("DataLayout", { DataLayout::NCHW })),
+                               framework::dataset::make("DataType", { DataType::F32 })))
 {
     // Validate output
     validate(CLAccessor(_target), _reference, tolerance_f32);
diff --git a/tests/validation/reference/Winograd.cpp b/tests/validation/reference/Winograd.cpp
index 604e252..757a06d 100644
--- a/tests/validation/reference/Winograd.cpp
+++ b/tests/validation/reference/Winograd.cpp
@@ -43,9 +43,6 @@
 template <typename T>
 void initialize_matrix_transform(SimpleTensor<T> &src, const Size2D &output_tile_size, const Size2D &kernel_size, WinogradTransformType winograd_transform_type)
 {
-    ARM_COMPUTE_ERROR_ON((output_tile_size != Size2D(2U, 2U)) && (output_tile_size != Size2D(4U, 4U)));
-    ARM_COMPUTE_ERROR_ON(kernel_size != Size2D(3U, 3U));
-
     // Winograd input transform matrices
     static const float imatrix2x2_3x3[] =
     {
@@ -86,6 +83,19 @@
         0.0f, 0.0f, 1.0f
     };
 
+    static const float fmatrix4x4_5x5[] =
+    {
+        1.0f, 0.0f, 0.0f, 0.0f, 0.0f,
+        -2.0f / 9.0f, -2.0f / 9.0f, -2.0f / 9.0f, -2.0f / 9.0f, -2.0f / 9.0f,
+        -2.0f / 9.0f, 2.0f / 9.0f, -2.0f / 9.0f, 2.0f / 9.0f, -2.0f / 9.0f,
+        1.0f / 90.0f, 1.0f / 45.0f, 2.0f / 45.0f, 4.0f / 45.0f, 8.0f / 45.0f,
+        1.0f / 90.0f, -1.0f / 45.0f, 2.0f / 45.0f, -4.0f / 45.0f, 8.0f / 45.0f,
+        4.0f / 45.0f, 2.0f / 45.0f, 1.0f / 45.0f, 1.0f / 90.0f, 1.0f / 180.0f,
+        4.0f / 45.0f, -2.0f / 45.0f, 1.0f / 45.0f, -1.0f / 90.0f, 1.0f / 180.0f,
+        0.0f, 0.0f, 0.0f, 0.0f, 1.0f
+
+    };
+
     // ------------------------------------------
 
     // Winograd output transform matrices
@@ -114,11 +124,12 @@
         { WinogradKey(std::pair<int, int>(4, 4), std::pair<int, int>(3, 3), WinogradTransformType::INPUT), imatrix4x4_3x3 },
         { WinogradKey(std::pair<int, int>(2, 2), std::pair<int, int>(3, 3), WinogradTransformType::FILTER), fmatrix2x2_3x3 },
         { WinogradKey(std::pair<int, int>(4, 4), std::pair<int, int>(3, 3), WinogradTransformType::FILTER), fmatrix4x4_3x3 },
+        { WinogradKey(std::pair<int, int>(4, 4), std::pair<int, int>(5, 5), WinogradTransformType::FILTER), fmatrix4x4_5x5 },
         { WinogradKey(std::pair<int, int>(2, 2), std::pair<int, int>(3, 3), WinogradTransformType::OUTPUT), omatrix2x2_3x3 },
         { WinogradKey(std::pair<int, int>(4, 4), std::pair<int, int>(3, 3), WinogradTransformType::OUTPUT), omatrix4x4_3x3 },
     };
 
-    // Find input matrix transform
+    // Find transformation matrix
     std::map<WinogradKey, const float *>::iterator it;
 
     it = matrix_map.find(WinogradKey(std::pair<int, int>(output_tile_size.width, output_tile_size.height),