diff --git a/tools/accuracy_checker/accuracy_checker/launcher/pytorch_launcher.py b/tools/accuracy_checker/accuracy_checker/launcher/pytorch_launcher.py
index cbb931b6eb9..fb413a323fe 100644
--- a/tools/accuracy_checker/accuracy_checker/launcher/pytorch_launcher.py
+++ b/tools/accuracy_checker/accuracy_checker/launcher/pytorch_launcher.py
@@ -20,7 +20,7 @@
 from collections import OrderedDict
 
 import numpy as np
-from ..config import PathField, StringField, DictField, NumberField, ListField
+from ..config import PathField, StringField, DictField, NumberField, ListField, BoolField
 from .launcher import Launcher
 
 MODULE_REGEX = r'(?:\w+)(?:(?:.\w+)*)'
@@ -51,7 +51,9 @@ def parameters(cls):
             'batch': NumberField(value_type=int, min_value=1, optional=True, description="Batch size.", default=1),
             'output_names': ListField(
                 optional=True, value_type=str, description='output tensor names'
-            )
+            ),
+            'use_openvino_backend': BoolField(
+                optional=True, default=False, description='use torch.compile feature with openvino backend')
         })
         return parameters
 
@@ -66,6 +68,13 @@ def __init__(self, config_entry: dict, *args, **kwargs):
                 import_error.msg)) from import_error
         self._torch = torch
         self.validate_config(config_entry)
+        self.is_openvino_backend = config_entry.get('use_openvino_backend')
+        if self.is_openvino_backend:
+            try:
+                import openvino.torch # pylint: disable=C0415, W0611
+            except ImportError as import_error:
+                raise ValueError("torch.compile is supported from OpenVINO 2023.1\n{}".format(
+                    import_error.msg)) from import_error
         module_args = config_entry.get("module_args", ())
         module_kwargs = config_entry.get("module_kwargs", {})
         self.device = self.get_value_from_config('device')
@@ -123,6 +132,9 @@ def load_module(self, model_cls, module_args, module_kwargs, checkpoint=None, st
                 module.load_state_dict(state, strict=False)
             module.to(self.device)
             module.eval()
+            if self.is_openvino_backend:
+                opts = {"device" : f"{self.device}"}
+                module = self._torch.compile(module, backend="openvino", options=opts)
             return module
 
     def fit_to_input(self, data, layer_name, layout, precision, template=None):
diff --git a/tools/accuracy_checker/accuracy_checker/launcher/pytorch_launcher_readme.md b/tools/accuracy_checker/accuracy_checker/launcher/pytorch_launcher_readme.md
index 795dc7bc4e3..232398a0066 100644
--- a/tools/accuracy_checker/accuracy_checker/launcher/pytorch_launcher_readme.md
+++ b/tools/accuracy_checker/accuracy_checker/launcher/pytorch_launcher_readme.md
@@ -12,6 +12,7 @@ For enabling PyTorch launcher you need to add `framework: pytorch` in launchers
 * `module_kwargs` - dictionary (`key`: `value` where `key` is argument name, `value` is argument value) which represent network module keyword arguments.
 * `adapter` - approach how raw output will be converted to representation of dataset problem, some adapters can be specific to framework. You can find detailed instruction how to use adapters [here](../adapters/README.md).
 * `batch` - batch size for running model (Optional, default 1).
+* `use_openvino_backend` - use torch.compile feature with `openvino` backend (Optional, default `False`)
 
 In turn if you model has several inputs you need to specify them in config, using specific parameter: `inputs`.
 Each input description should has following info: