diff --git a/.github/workflows/docs.yml b/.github/workflows/docs.yml
index 286f665a..eb5788ce 100644
--- a/.github/workflows/docs.yml
+++ b/.github/workflows/docs.yml
@@ -9,10 +9,10 @@ jobs:
   docs:
     runs-on: ubuntu-latest
     steps:
-    - uses: actions/checkout@v2
+    - uses: actions/checkout@v1
 
     - name: Python
-      uses: actions/setup-python@v1
+      uses: actions/setup-python@v2
       with:
         python-version: 3.8
 
diff --git a/.github/workflows/tests.yml b/.github/workflows/tests.yml
index dd53134b..9fe6deca 100644
--- a/.github/workflows/tests.yml
+++ b/.github/workflows/tests.yml
@@ -78,8 +78,27 @@ jobs:
         python-version: ${{ matrix.python-version }}
     - name: Install package and dependencies
       run: pip install invoke .[test]
-    - name: invoke pytest
-      run: invoke pytest
+    - name: invoke unit
+      run: invoke unit
+
+
+  unit-pretrained:
+    runs-on: ${{ matrix.os }}
+    strategy:
+      matrix:
+        python-version: ['3.11']
+        os: [ubuntu-latest, macos-latest, windows-latest]
+    steps:
+    - uses: actions/checkout@v1
+    - name: Set up Python ${{ matrix.python-version }}
+      uses: actions/setup-python@v2
+      with:
+        python-version: ${{ matrix.python-version }}
+    - name: Install package and dependencies
+      run:  |
+          pip install invoke pytest .[pretrained]
+    - name: invoke pretrained
+      run: invoke pretrained
 
 
   minimum:
@@ -125,7 +144,7 @@ jobs:
       run: invoke tutorials
 
 
-  pretrained:
+  pretrained-tutorials:
     runs-on: ${{ matrix.os }}
     strategy:
       matrix:
@@ -141,5 +160,5 @@ jobs:
       run:  |
           pip install "mistune>=2.0.3,<3.1"
           pip install invoke jupyter .[pretrained]
-    - name: invoke pretrained
-      run: invoke pretrained
+    - name: invoke pretrained-tutorials
+      run: invoke pretrained-tutorials
diff --git a/Makefile b/Makefile
index 47bcf204..6787fd92 100644
--- a/Makefile
+++ b/Makefile
@@ -125,7 +125,7 @@ fix-lint: ## fix lint issues using autoflake, autopep8, and isort
 
 .PHONY: test-unit
 test-unit: ## run tests quickly with the default Python
-	invoke pytest
+	invoke unit
 
 .PHONY: test-readme
 test-readme: ## run the readme snippets
@@ -135,9 +135,13 @@ test-readme: ## run the readme snippets
 test-tutorials: ## run the tutorial notebooks
 	invoke tutorials
 
+.PHONY: test-pretrained
+test-pretrained: ## run the tutorial notebooks
+	invoke pretrained
+
 .PHONY: test-pretrained-tutorials
 test-pretrained-tutorials: ## run the tutorial notebooks
-	invoke pretrained
+	invoke pretrained-tutorials
 
 .PHONY: test
 test: test-unit test-readme test-tutorials ## test everything that needs test dependencies
diff --git a/orion/primitives/jsons/orion.primitives.timesfm.TimesFM.json b/orion/primitives/jsons/orion.primitives.timesfm.TimesFM.json
index 42388102..77a4e965 100644
--- a/orion/primitives/jsons/orion.primitives.timesfm.TimesFM.json
+++ b/orion/primitives/jsons/orion.primitives.timesfm.TimesFM.json
@@ -17,6 +17,11 @@
             {
                 "name": "X",
                 "type": "ndarray"
+            },
+            {
+                "name": "force",
+                "type": "bool",
+                "default": false
             }
         ],
         "output": [
diff --git a/orion/primitives/timesfm.py b/orion/primitives/timesfm.py
index defd5770..0ce23c52 100644
--- a/orion/primitives/timesfm.py
+++ b/orion/primitives/timesfm.py
@@ -8,9 +8,27 @@
 https://github.com/google-research/timesfm?tab=readme-ov-file
 """
 
-import numpy as np
+import sys
 
-import timesfm as tf
+if sys.version_info < (3, 11):
+    msg = (
+        '`timesfm` requires Python >= 3.11 and your '
+        f'python version is {sys.version}.\n'
+        'Make sure you are using Python 3.11 or later.\n'
+    )
+    raise RuntimeError(msg)
+
+try:
+    import timesfm as tf
+except ImportError as ie:
+    ie.msg += (
+        '\n\nIt seems like `timesfm` is not installed.\n'
+        'Please install `timesfm` using:\n'
+        '\n    pip install orion-ml[pretrained]'
+    )
+    raise
+
+MAX_LENGTH = 93000
 
 
 class TimesFM:
@@ -53,7 +71,7 @@ def __init__(self,
                                                           horizon_len=pred_len),
                                 checkpoint=tf.TimesFmCheckpoint(huggingface_repo_id=repo_id))
 
-    def predict(self, X):
+    def predict(self, X, force=False):
         """Forecasting timeseries
 
         Args:
@@ -63,30 +81,21 @@ def predict(self, X):
             ndarray:
                 forecasted timeseries.
         """
-        frequency_input = [self.freq]*len(X)
-        d = X.shape[-1]
-
-        # Univariate
-        if d == 1:
-            y_hat, _ = self.model.forecast(X[:, :, 0], freq=frequency_input)
-            return y_hat[:, 0]
-
-        # Multivariate
-        covariates = list(range(d))
-        covariates = covariates.remove(self.target)
-        X_cont = X[:, :, self.target]
-        X_cov = np.delete(X, self.target, axis=2)
-
-        # Append covariates with future values
-        m, n, k = X_cov.shape
-        X_cov_new = np.zeros((m, n+self.pred_len, k))
-        X_cov_new[:, :-self.pred_len, :] = X_cov
-        X_cov_new[:-1, -self.pred_len:, :] = X_cov[1:, :self.pred_len, :]
-
-        x_cov = {str(i): X_cov_new[:, :, i] for i in range(k)}
-        y_hat, _ = self.model.forecast_with_covariates(
-            inputs=X_cont,
-            dynamic_numerical_covariates=x_cov,
-            freq=frequency_input,
-        )
-        return np.concatenate(y_hat)
+        frequency_input = [self.freq] * len(X)
+        m, n, d = X.shape
+
+        # does not support multivariate
+        if d > 1:
+            raise ValueError(f'Encountered X with too many channels (channels={d}).')
+
+        # does not support long time series
+        if not force and m > (MAX_LENGTH - self.window_size):
+            msg = (
+                f'`X` has {m} samples, which might result in out of memory issues.\n'
+                'If you are sure you want to proceed, set `force=True`.'
+            )
+
+            raise MemoryError(msg)
+
+        y_hat, _ = self.model.forecast(X[:, :, 0], freq=frequency_input)
+        return y_hat[:, 0]
diff --git a/setup.py b/setup.py
index 166dbf46..97d6e1e3 100644
--- a/setup.py
+++ b/setup.py
@@ -133,7 +133,6 @@
         'test': tests_require,
         'dev': development_requires + tests_require,
         'pretrained': pretrained_requires,
-        'pretrained-dev': pretrained_requires + development_requires + tests_require,
     },
     include_package_data=True,
     install_requires=install_requires,
diff --git a/tasks.py b/tasks.py
index 0c98d677..9bf691c0 100644
--- a/tasks.py
+++ b/tasks.py
@@ -19,8 +19,13 @@
 
 
 @task
-def pytest(c):
-    c.run('python -m pytest --cov=orion')
+def unit(c):
+    c.run('python -m pytest ./tests/unit --cov=orion')
+
+
+@task
+def pretrained(c):
+    c.run('python -m pytest ./tests/pretrained')
 
 
 @task
@@ -70,7 +75,7 @@ def install_minimum(c):
 def minimum(c):
     install_minimum(c)
     c.run('python -m pip check')
-    c.run('python -m pytest')
+    c.run('python -m pytest ./tests/unit')
 
 
 @task
@@ -107,7 +112,7 @@ def tutorials(c):
             ), hide='out')
 
 @task
-def pretrained(c):
+def pretrained_tutorials(c):
     pipelines = os.listdir(os.path.join('orion', 'pipelines', 'pretrained'))
     for ipynb_file in glob.glob('tutorials/pipelines/*.ipynb'):
         for pipeline in pipelines: 
diff --git a/tests/pretrained/test_timesfm.py b/tests/pretrained/test_timesfm.py
new file mode 100644
index 00000000..6f5668e8
--- /dev/null
+++ b/tests/pretrained/test_timesfm.py
@@ -0,0 +1,66 @@
+import importlib
+import unittest
+from unittest.mock import patch
+
+import numpy as np
+
+import timesfm as tf
+from orion.primitives.timesfm import MAX_LENGTH, TimesFM
+
+
+class TestTimesFMImport(unittest.TestCase):
+
+    @patch('sys.version_info', (3, 10))
+    def test_runtime_error_python_version_less_than_3_11(self):
+        with self.assertRaises(RuntimeError) as context:
+            import orion.primitives.timesfm
+            importlib.reload(orion.primitives.timesfm)
+
+        self.assertIn('requires Python >= 3.11', str(context.exception))
+        self.assertIn('python version is', str(context.exception))
+
+    @patch('sys.version_info', (3, 11))
+    @patch('builtins.__import__', side_effect=ImportError())
+    def test_import_error_timesfm_not_installed(self, mock_import):
+        # simulate Python version 3.11 and timesfm not installed
+        with self.assertRaises(ImportError):
+            import orion.primitives.timesfm  # noqa
+
+
+class TestTimesFMPredict(unittest.TestCase):
+
+    def setUp(self):
+        self.model = TimesFM()
+
+    def test_value_error_multivariate_input(self):
+        # create a multivariate input with more than one channel
+        X = np.random.rand(10, 5, 2)  # Shape (m, n, d) with d > 1
+
+        with self.assertRaises(ValueError) as context:
+            self.model.predict(X)
+
+        self.assertIn('Encountered X with too many channels', str(context.exception))
+
+    def test_memory_error_long_time_series(self):
+        # create a long time series input
+        m = MAX_LENGTH - self.model.window_size + 1
+        X = np.random.rand(m, 5, 1)  # Shape (m, n, d) with d = 1
+
+        with self.assertRaises(MemoryError) as context:
+            self.model.predict(X)
+
+        self.assertIn('might result in out of memory issues', str(context.exception))
+
+    @patch.object(tf.TimesFm, 'forecast', return_value=(np.random.rand(10, 1), None))
+    def test_no_memory_error_with_force(self, mock_forecast):
+        # create a long time series input
+        m = MAX_LENGTH - self.model.window_size + 1
+        X = np.random.rand(m, 5, 1)  # Shape (m, n, d) with d = 1
+
+        # should not raise MemoryError when force=True
+        try:
+            self.model.predict(X, force=True)
+        except MemoryError:
+            self.fail("predict() raised MemoryError unexpectedly with force=True")
+
+        mock_forecast.assert_called_once()
diff --git a/tests/evaluation/test_common.py b/tests/unit/evaluation/test_common.py
similarity index 100%
rename from tests/evaluation/test_common.py
rename to tests/unit/evaluation/test_common.py
diff --git a/tests/evaluation/test_contextual.py b/tests/unit/evaluation/test_contextual.py
similarity index 100%
rename from tests/evaluation/test_contextual.py
rename to tests/unit/evaluation/test_contextual.py
diff --git a/tests/evaluation/test_point.py b/tests/unit/evaluation/test_point.py
similarity index 100%
rename from tests/evaluation/test_point.py
rename to tests/unit/evaluation/test_point.py
diff --git a/tests/evaluation/test_utils.py b/tests/unit/evaluation/test_utils.py
similarity index 100%
rename from tests/evaluation/test_utils.py
rename to tests/unit/evaluation/test_utils.py
diff --git a/tests/primitives/adapters/test_ncps.py b/tests/unit/primitives/adapters/test_ncps.py
similarity index 100%
rename from tests/primitives/adapters/test_ncps.py
rename to tests/unit/primitives/adapters/test_ncps.py
diff --git a/tests/primitives/test_anomaly_transformer.py b/tests/unit/primitives/test_anomaly_transformer.py
similarity index 100%
rename from tests/primitives/test_anomaly_transformer.py
rename to tests/unit/primitives/test_anomaly_transformer.py
diff --git a/tests/primitives/test_timeseries_anomalies.py b/tests/unit/primitives/test_timeseries_anomalies.py
similarity index 100%
rename from tests/primitives/test_timeseries_anomalies.py
rename to tests/unit/primitives/test_timeseries_anomalies.py
diff --git a/tests/primitives/test_timeseries_errors.py b/tests/unit/primitives/test_timeseries_errors.py
similarity index 100%
rename from tests/primitives/test_timeseries_errors.py
rename to tests/unit/primitives/test_timeseries_errors.py
diff --git a/tests/primitives/test_timeseries_preprocessing.py b/tests/unit/primitives/test_timeseries_preprocessing.py
similarity index 100%
rename from tests/primitives/test_timeseries_preprocessing.py
rename to tests/unit/primitives/test_timeseries_preprocessing.py
diff --git a/tests/test_analysis.py b/tests/unit/test_analysis.py
similarity index 100%
rename from tests/test_analysis.py
rename to tests/unit/test_analysis.py
diff --git a/tests/test_benchmark.py b/tests/unit/test_benchmark.py
similarity index 100%
rename from tests/test_benchmark.py
rename to tests/unit/test_benchmark.py
diff --git a/tests/test_core.py b/tests/unit/test_core.py
similarity index 100%
rename from tests/test_core.py
rename to tests/unit/test_core.py
diff --git a/tests/test_data.py b/tests/unit/test_data.py
similarity index 100%
rename from tests/test_data.py
rename to tests/unit/test_data.py
diff --git a/tests/test_functional.py b/tests/unit/test_functional.py
similarity index 100%
rename from tests/test_functional.py
rename to tests/unit/test_functional.py