diff --git a/test/distributed/test_nccl.py b/test/distributed/test_nccl.py
index a90288bd05f..6991b77a4ee 100644
--- a/test/distributed/test_nccl.py
+++ b/test/distributed/test_nccl.py
@@ -13,6 +13,7 @@ from torch.testing._internal.common_utils import (
     load_tests,
     TEST_WITH_ROCM,
     skip_but_pass_in_sandcastle_if,
+    NoTest,
 )
 from torch.testing._internal.common_cuda import TEST_CUDA, TEST_MULTIGPU
 from torch.testing._internal.common_device_type import (
@@ -34,7 +35,7 @@ load_tests = load_tests
 nGPUs = torch.cuda.device_count()
 if not TEST_CUDA:
     print("CUDA not available, skipping tests", file=sys.stderr)
-    TestCase = object  # noqa: F811
+    TestCase = NoTest  # noqa: F811
 
 
 datatypes = [torch.float]
diff --git a/test/jit/test_cuda.py b/test/jit/test_cuda.py
index 6937af9f292..40603a73462 100644
--- a/test/jit/test_cuda.py
+++ b/test/jit/test_cuda.py
@@ -9,7 +9,7 @@ import torch
 from typing import NamedTuple
 from torch.testing import FileCheck
 from torch.testing._internal.jit_utils import JitTestCase
-from torch.testing._internal.common_utils import skipIfRocm, skipCUDANonDefaultStreamIf
+from torch.testing._internal.common_utils import skipIfRocm, skipCUDANonDefaultStreamIf, NoTest
 
 # Make the helper files in test/ importable
 pytorch_test_dir = os.path.dirname(os.path.dirname(os.path.realpath(__file__)))
@@ -23,7 +23,7 @@ TEST_MULTIGPU = TEST_CUDA and torch.cuda.device_count() >= 2
 # If GPU is not available, then do not run the tests
 if not TEST_CUDA:
     print('CUDA not available, skipping tests', file=sys.stderr)
-    JitTestCase = object  # noqa: F811
+    JitTestCase = NoTest  # noqa: F811
 
 TEST_LARGE_TENSOR = TEST_CUDA
 
diff --git a/test/jit/test_pdt.py b/test/jit/test_pdt.py
index dd8c0068511..5fa39b8cac3 100644
--- a/test/jit/test_pdt.py
+++ b/test/jit/test_pdt.py
@@ -6,6 +6,7 @@ import torch
 from torch.testing._internal.jit_utils import JitTestCase, make_global
 from torch.jit._monkeytype_config import _IS_MONKEYTYPE_INSTALLED
 from typing import List, Dict, Tuple, Any, Optional, NamedTuple  # noqa: F401
+from torch.testing._internal.common_utils import NoTest
 
 # Make the helper files in test/ importable
 pytorch_test_dir = os.path.dirname(os.path.dirname(os.path.realpath(__file__)))
@@ -13,7 +14,7 @@ sys.path.append(pytorch_test_dir)
 
 if not _IS_MONKEYTYPE_INSTALLED:
     print("monkeytype is not installed. Skipping tests for Profile-Directed Typing", file=sys.stderr)
-    JitTestCase = object  # type: ignore[misc, assignment] # noqa: F811
+    JitTestCase = NoTest  # type: ignore[misc, assignment] # noqa: F811
 
 if __name__ == "__main__":
     raise RuntimeError(
diff --git a/test/run_test.py b/test/run_test.py
index 708d22c9a37..340e1b17023 100755
--- a/test/run_test.py
+++ b/test/run_test.py
@@ -887,21 +887,10 @@ CUSTOM_HANDLERS = {
 
 PYTEST_BLOCKLIST = [
     "test_package",
-    "test_nccl",
     "inductor/test_torchinductor",
-    "test_cuda",
     "test_quantization",
-    "test_cuda_nvml_based_avail",
-    "test_cuda_primary_ctx",
-    "test_cuda_sanitizer",
-    "test_cuda_trace",
     "test_fx",
-    "test_jiterator",
-    "test_mps",
-    "test_cuda_trace",
     "profiler/test_profiler",
-    "test_jit",
-    "test_jit_legacy",
     "dynamo/test_repros",  # skip_if_pytest
     "dynamo/test_optimizers",  # skip_if_pytest
     "dynamo/test_dynamic_shapes",  # needs change to check_if_enable for disabled test issues
diff --git a/test/test_cuda.py b/test/test_cuda.py
index 92d68f92281..5fcbdabc87d 100644
--- a/test/test_cuda.py
+++ b/test/test_cuda.py
@@ -28,7 +28,7 @@ from torch.utils.checkpoint import checkpoint_sequential
 from torch.testing._internal.common_utils import TestCase, freeze_rng_state, run_tests, \
     NO_MULTIPROCESSING_SPAWN, skipIfRocm, load_tests, IS_REMOTE_GPU, IS_SANDCASTLE, IS_WINDOWS, \
     slowTest, skipCUDANonDefaultStreamIf, skipCUDAMemoryLeakCheckIf, TEST_WITH_ROCM, TEST_NUMPY, \
-    get_cycles_per_ms, parametrize, instantiate_parametrized_tests, subtest, IS_JETSON, gcIfJetson
+    get_cycles_per_ms, parametrize, instantiate_parametrized_tests, subtest, IS_JETSON, gcIfJetson, NoTest
 from torch.testing._internal.autocast_test_lists import AutocastTestLists
 
 # load_tests from common_utils is used to automatically filter tests for
@@ -44,7 +44,7 @@ TEST_MULTIGPU = TEST_CUDA and torch.cuda.device_count() >= 2
 
 if not TEST_CUDA:
     print('CUDA not available, skipping tests', file=sys.stderr)
-    TestCase = object  # noqa: F811
+    TestCase = NoTest  # noqa: F811
 
 TEST_CUDAMALLOCASYNC = TEST_CUDA and (torch.cuda.get_allocator_backend() == "cudaMallocAsync")
 TEST_LARGE_TENSOR = TEST_CUDA
diff --git a/test/test_cuda_nvml_based_avail.py b/test/test_cuda_nvml_based_avail.py
index 04bad0ff86a..012c047563e 100644
--- a/test/test_cuda_nvml_based_avail.py
+++ b/test/test_cuda_nvml_based_avail.py
@@ -13,7 +13,7 @@ with patch.dict(os.environ, {"PYTORCH_NVML_BASED_CUDA_CHECK": "1"}):
     # Before executing the desired tests, we need to disable CUDA initialization and fork_handler additions that would
     # otherwise be triggered by the `torch.testing._internal.common_utils` module import
     from torch.testing._internal.common_utils import (parametrize, instantiate_parametrized_tests, run_tests, TestCase,
-                                                      IS_WINDOWS, IS_JETSON)
+                                                      IS_WINDOWS, IS_JETSON, NoTest)
     # NOTE: Because `remove_device_and_dtype_suffixes` initializes CUDA context (triggered via the import of
     # `torch.testing._internal.common_device_type` which imports `torch.testing._internal.common_cuda`) we need
     # to bypass that method here which should be irrelevant to the parameterized tests in this module.
@@ -22,7 +22,7 @@ with patch.dict(os.environ, {"PYTORCH_NVML_BASED_CUDA_CHECK": "1"}):
     TEST_CUDA = torch.cuda.is_available()
     if not TEST_CUDA:
         print('CUDA not available, skipping tests', file=sys.stderr)
-        TestCase = object  # type: ignore[misc, assignment] # noqa: F811
+        TestCase = NoTest  # type: ignore[misc, assignment] # noqa: F811
 
 
 class TestExtendedCUDAIsAvail(TestCase):
diff --git a/test/test_cuda_primary_ctx.py b/test/test_cuda_primary_ctx.py
index 74005515a49..b0fa0e14c79 100644
--- a/test/test_cuda_primary_ctx.py
+++ b/test/test_cuda_primary_ctx.py
@@ -1,7 +1,7 @@
 # Owner(s): ["module: cuda"]
 
 import torch
-from torch.testing._internal.common_utils import TestCase, run_tests, skipIfRocmVersionLessThan
+from torch.testing._internal.common_utils import TestCase, run_tests, skipIfRocmVersionLessThan, NoTest
 import sys
 import unittest
 
@@ -16,7 +16,7 @@ TEST_MULTIGPU = TEST_CUDA and torch.cuda.device_count() >= 2
 
 if not TEST_CUDA:
     print('CUDA not available, skipping tests', file=sys.stderr)
-    TestCase = object  # noqa: F811
+    TestCase = NoTest  # noqa: F811
 
 
 class TestCudaPrimaryCtx(TestCase):
diff --git a/test/test_cuda_sanitizer.py b/test/test_cuda_sanitizer.py
index f8733ba43a4..b52fafb94ae 100644
--- a/test/test_cuda_sanitizer.py
+++ b/test/test_cuda_sanitizer.py
@@ -8,7 +8,7 @@ from typing import List
 import torch
 import torch.cuda._sanitizer as csan
 from torch.cuda._sanitizer import StreamId, DataPtr, EventId
-from torch.testing._internal.common_utils import TestCase, run_tests
+from torch.testing._internal.common_utils import TestCase, run_tests, NoTest
 
 
 # We cannot import TEST_CUDA from torch.testing._internal.common_cuda here,
@@ -19,7 +19,7 @@ TEST_CUDA = torch.cuda.is_available()
 
 if not TEST_CUDA:
     print("CUDA not available, skipping tests", file=sys.stderr)
-    TestCase = object  # noqa: F811
+    TestCase = NoTest  # noqa: F811
 
 
 class TestArgumentHandler(TestCase):
diff --git a/test/test_cuda_trace.py b/test/test_cuda_trace.py
index 07ba30d27f4..6581ffffec0 100644
--- a/test/test_cuda_trace.py
+++ b/test/test_cuda_trace.py
@@ -6,7 +6,7 @@ import unittest.mock
 
 import torch
 import torch.utils._cuda_trace as cuda_trace
-from torch.testing._internal.common_utils import TestCase, run_tests
+from torch.testing._internal.common_utils import TestCase, run_tests, NoTest
 
 # NOTE: Each test needs to be run in a brand new process, to reset the registered hooks
 # and make sure the CUDA streams are initialized for each test that uses them.
@@ -19,7 +19,7 @@ TEST_CUDA = torch.cuda.is_available()
 
 if not TEST_CUDA:
     print("CUDA not available, skipping tests", file=sys.stderr)
-    TestCase = object  # noqa: F811
+    TestCase = NoTest  # noqa: F811
 
 
 class TestCudaTrace(TestCase):
diff --git a/test/test_jiterator.py b/test/test_jiterator.py
index f995e540887..1acba982f3a 100644
--- a/test/test_jiterator.py
+++ b/test/test_jiterator.py
@@ -5,7 +5,7 @@ from torch.cuda.jiterator import _create_jit_fn as create_jit_fn
 from torch.cuda.jiterator import _create_multi_output_jit_fn as create_multi_output_jit_fn
 import sys
 from itertools import product
-from torch.testing._internal.common_utils import TestCase, parametrize, run_tests, TEST_CUDA
+from torch.testing._internal.common_utils import TestCase, parametrize, run_tests, TEST_CUDA, NoTest
 from torch.testing._internal.common_dtype import all_types_and_complex_and
 from torch.testing._internal.common_device_type import (
     skipCUDAIfRocm, skipCUDAIf, instantiate_device_type_tests, dtypes, toleranceOverride, tol)
@@ -13,7 +13,7 @@ from torch.testing._internal.common_cuda import _get_torch_cuda_version
 
 if not TEST_CUDA:
     print('CUDA not available, skipping tests', file=sys.stderr)
-    TestCase = object  # noqa: F811
+    TestCase = NoTest  # noqa: F811
 
 
 code_string = "template <typename T> T my_fused_kernel(T x, T y, T alpha, T beta) { return alpha * x + beta * y; }"
diff --git a/test/test_mps.py b/test/test_mps.py
index 4e5e42d3b22..6abfb45306d 100644
--- a/test/test_mps.py
+++ b/test/test_mps.py
@@ -22,7 +22,7 @@ from torch import inf
 from torch.nn import Parameter
 from torch.testing._internal import opinfo
 from torch.testing._internal.common_utils import \
-    (gradcheck, gradgradcheck, run_tests, TestCase, download_file, IS_CI,
+    (gradcheck, gradgradcheck, run_tests, TestCase, download_file, IS_CI, NoTest,
      TEST_WITH_UBSAN, dtype_abbrs, skipIfSlowGradcheckEnv, TEST_WITH_ASAN, suppress_warnings)
 from torch.testing import make_tensor
 from torch.testing._comparison import TensorLikePair
@@ -175,8 +175,8 @@ def mps_ops_modifier(ops):
 # Same logic as test_cuda.py
 if not torch.backends.mps.is_available():
     print('MPS not available, skipping tests', file=sys.stderr)
-    TestCase = object  # noqa: F811
-    NNTestCase = object  # noqa: F811
+    TestCase = NoTest  # noqa: F811
+    NNTestCase = NoTest  # noqa: F811
 
 product_version = float('.'.join(platform.mac_ver()[0].split('.')[:2]))
 
diff --git a/torch/testing/_internal/common_utils.py b/torch/testing/_internal/common_utils.py
index 55d6200eb1d..78569f6b9b3 100644
--- a/torch/testing/_internal/common_utils.py
+++ b/torch/testing/_internal/common_utils.py
@@ -1686,8 +1686,6 @@ def remove_device_and_dtype_suffixes(test_name: str) -> str:
 
 def check_if_enable(test: unittest.TestCase):
     test_suite = str(test.__class__).split('\'')[1]
-    if "USING_PYTEST" in os.environ:
-        test_suite = f"__main__.{test_suite.split('.')[1]}"
     raw_test_name = f'{test._testMethodName} ({test_suite})'
     if raw_test_name in slow_tests_dict:
         getattr(test, test._testMethodName).__dict__['slow_test'] = True
@@ -1976,6 +1974,11 @@ def set_warn_always_context(new_val: bool):
         torch.set_warn_always(old_val)
 
 
+class NoTest():
+    # causes pytest to not recognize this class as a test
+    __test__ = False
+
+
 class TestCase(expecttest.TestCase):
     # NOTE: "precision" lets classes and generated tests set minimum
     # atol values when comparing tensors. Used by @precisionOverride and @toleranceOverride, for