diff --git a/setup.py b/setup.py
index 7ef3652f87..0c4bbc1541 100644
--- a/setup.py
+++ b/setup.py
@@ -62,7 +62,7 @@
 
 EXTRAS_REQUIRE = {
     "nncf": ["nncf>=2.11.0"],
-    "openvino": ["nncf>=2.11.0", "openvino==2024.4.1.dev20240926", "openvino-tokenizers==2024.4.1.0.dev20240926"],
+    "openvino": ["nncf>=2.11.0", "openvino==2024.5.0", "openvino-tokenizers==2024.5.0"],
     "neural-compressor": ["neural-compressor[pt]>3.0", "accelerate", "transformers<4.46"],
     "ipex": ["intel-extension-for-pytorch", "transformers>=4.39,<4.45"],
     "diffusers": ["diffusers"],
diff --git a/tests/openvino/test_exporters_cli.py b/tests/openvino/test_exporters_cli.py
index f218fa05ba..8ce707e142 100644
--- a/tests/openvino/test_exporters_cli.py
+++ b/tests/openvino/test_exporters_cli.py
@@ -50,6 +50,7 @@
 from optimum.intel.utils.import_utils import (
     compare_versions,
     is_openvino_tokenizers_available,
+    is_openvino_version,
     is_tokenizers_version,
     is_transformers_version,
 )
@@ -80,20 +81,20 @@ class OVCLIExportTestCase(unittest.TestCase):
     if is_transformers_version(">=", "4.45"):
         SUPPORTED_ARCHITECTURES.extend([("text-to-image", "stable-diffusion-3"), ("text-to-image", "flux")])
     EXPECTED_NUMBER_OF_TOKENIZER_MODELS = {
-        "gpt2": 2 if is_tokenizers_version("<", "0.20") else 0,
+        "gpt2": 2 if is_tokenizers_version("<", "0.20") or is_openvino_version(">=", "2024.5") else 0,
         "t5": 0,  # no .model file in the repository
         "albert": 0,  # not supported yet
         "distilbert": 1,  # no detokenizer
-        "roberta": 2 if is_tokenizers_version("<", "0.20") else 0,
+        "roberta": 2 if is_tokenizers_version("<", "0.20") or is_openvino_version(">=", "2024.5") else 0,
         "vit": 0,  # no tokenizer for image model
         "wav2vec2": 0,  # no tokenizer
         "bert": 1,  # no detokenizer
-        "blenderbot": 2 if is_tokenizers_version("<", "0.20") else 0,
-        "stable-diffusion": 2 if is_tokenizers_version("<", "0.20") else 0,
-        "stable-diffusion-xl": 4 if is_tokenizers_version("<", "0.20") else 0,
-        "stable-diffusion-3": 6 if is_tokenizers_version("<", "0.20") else 2,
-        "flux": 4 if is_tokenizers_version("<", "0.20") else 0,
-        "llava": 2 if is_tokenizers_version("<", "0.20") else 0,
+        "blenderbot": 2 if is_tokenizers_version("<", "0.20") or is_openvino_version(">=", "2024.5") else 0,
+        "stable-diffusion": 2 if is_tokenizers_version("<", "0.20") or is_openvino_version(">=", "2024.5") else 0,
+        "stable-diffusion-xl": 4 if is_tokenizers_version("<", "0.20") or is_openvino_version(">=", "2024.5") else 0,
+        "stable-diffusion-3": 6 if is_tokenizers_version("<", "0.20") or is_openvino_version(">=", "2024.5") else 2,
+        "flux": 4 if is_tokenizers_version("<", "0.20") or is_openvino_version(">=", "2024.5") else 0,
+        "llava": 2 if is_tokenizers_version("<", "0.20") or is_openvino_version(">=", "2024.5") else 0,
     }
 
     SUPPORTED_SD_HYBRID_ARCHITECTURES = [
diff --git a/tests/openvino/test_quantization.py b/tests/openvino/test_quantization.py
index 7a415c3a3a..c8ce4be89a 100644
--- a/tests/openvino/test_quantization.py
+++ b/tests/openvino/test_quantization.py
@@ -258,7 +258,7 @@ class OVWeightCompressionTest(unittest.TestCase):
                 sensitivity_metric="mean_activation_magnitude",
                 dataset=["one two, " * i for i in range(10)],
             ),
-            {"int4": 25, "int8": 14},
+            {"int4": 24, "int8": 16},
         ),
         (
             OVModelForCausalLM,
@@ -883,7 +883,7 @@ def preprocess_function(examples, tokenizer):
 
 
 class OVTrainerTest(unittest.TestCase):
-    SUPPORTED_ARCHITECTURES_WITH_EXPECTED_QUANTIZED_MATMULS = (("albert", 64, 39),)
+    SUPPORTED_ARCHITECTURES_WITH_EXPECTED_QUANTIZED_MATMULS = (("albert", 63, 39),)
 
     @parameterized.expand(SUPPORTED_ARCHITECTURES_WITH_EXPECTED_QUANTIZED_MATMULS)
     @unittest.skipIf(