diff --git a/cirq-core/cirq/devices/noise_utils.py b/cirq-core/cirq/devices/noise_utils.py
index 7fb30a040ba..917f873dace 100644
--- a/cirq-core/cirq/devices/noise_utils.py
+++ b/cirq-core/cirq/devices/noise_utils.py
@@ -13,10 +13,9 @@
 # limitations under the License.
 
 from typing import TYPE_CHECKING, Any, Dict, Tuple, Type, Union
-import numpy as np
 
-from cirq import ops, protocols, value
-from cirq._compat import proper_repr
+from cirq import ops, protocols, value, qis
+from cirq._compat import proper_repr, deprecated
 
 if TYPE_CHECKING:
     import cirq
@@ -97,8 +96,10 @@ def _from_json_dict_(cls, gate_type, qubits, **kwargs) -> 'OpIdentifier':
         return cls(gate_type, *qubits)
 
 
-# TODO: expose all from top-level cirq?
-def decay_constant_to_xeb_fidelity(decay_constant: float, num_qubits: int = 2) -> float:
+@deprecated(deadline='v2.0', fix='use cirq.qis.decay_constant_to_xeb_fidelity')
+def decay_constant_to_xeb_fidelity(
+    decay_constant: float, num_qubits: int = 2
+) -> float:  # pragma: no cover
     """Calculates the XEB fidelity from the depolarization decay constant.
 
     Args:
@@ -108,11 +109,13 @@ def decay_constant_to_xeb_fidelity(decay_constant: float, num_qubits: int = 2) -
     Returns:
         Calculated XEB fidelity.
     """
-    N = 2**num_qubits
-    return 1 - ((1 - decay_constant) * (1 - 1 / N))
+    return qis.decay_constant_to_xeb_fidelity(decay_constant, num_qubits)
 
 
-def decay_constant_to_pauli_error(decay_constant: float, num_qubits: int = 1) -> float:
+@deprecated(deadline='v2.0', fix='use cirq.qis.decay_constant_to_pauli_error')
+def decay_constant_to_pauli_error(
+    decay_constant: float, num_qubits: int = 1
+) -> float:  # pragma: no cover
     """Calculates pauli error from the depolarization decay constant.
 
     Args:
@@ -122,11 +125,13 @@ def decay_constant_to_pauli_error(decay_constant: float, num_qubits: int = 1) ->
     Returns:
         Calculated Pauli error.
     """
-    N = 2**num_qubits
-    return (1 - decay_constant) * (1 - 1 / N / N)
+    return qis.decay_constant_to_pauli_error(decay_constant, num_qubits)
 
 
-def pauli_error_to_decay_constant(pauli_error: float, num_qubits: int = 1) -> float:
+@deprecated(deadline='v2.0', fix='use cirq.qis.pauli_error_to_decay_constant')
+def pauli_error_to_decay_constant(
+    pauli_error: float, num_qubits: int = 1
+) -> float:  # pragma: no cover
     """Calculates depolarization decay constant from pauli error.
 
     Args:
@@ -136,11 +141,13 @@ def pauli_error_to_decay_constant(pauli_error: float, num_qubits: int = 1) -> fl
     Returns:
         Calculated depolarization decay constant.
     """
-    N = 2**num_qubits
-    return 1 - (pauli_error / (1 - 1 / N / N))
+    return qis.pauli_error_to_decay_constant(pauli_error, num_qubits)
 
 
-def xeb_fidelity_to_decay_constant(xeb_fidelity: float, num_qubits: int = 2) -> float:
+@deprecated(deadline='v2.0', fix='use cirq.qis.xeb_fidelity_to_decay_constant')
+def xeb_fidelity_to_decay_constant(
+    xeb_fidelity: float, num_qubits: int = 2
+) -> float:  # pragma: no cover
     """Calculates the depolarization decay constant from XEB fidelity.
 
     Args:
@@ -150,11 +157,11 @@ def xeb_fidelity_to_decay_constant(xeb_fidelity: float, num_qubits: int = 2) ->
     Returns:
         Calculated depolarization decay constant.
     """
-    N = 2**num_qubits
-    return 1 - (1 - xeb_fidelity) / (1 - 1 / N)
+    return qis.xeb_fidelity_to_decay_constant(xeb_fidelity, num_qubits)
 
 
-def pauli_error_from_t1(t_ns: float, t1_ns: float) -> float:
+@deprecated(deadline='v2.0', fix='use cirq.qis.pauli_error_from_t1')
+def pauli_error_from_t1(t_ns: float, t1_ns: float) -> float:  # pragma: no cover
     """Calculates the pauli error from T1 decay constant.
 
     This computes error for a specific duration, `t`.
@@ -166,11 +173,11 @@ def pauli_error_from_t1(t_ns: float, t1_ns: float) -> float:
     Returns:
         Calculated Pauli error resulting from T1 decay.
     """
-    t2 = 2 * t1_ns
-    return (1 - np.exp(-t_ns / t2)) / 2 + (1 - np.exp(-t_ns / t1_ns)) / 4
+    return qis.pauli_error_from_t1(t_ns, t1_ns)
 
 
-def average_error(decay_constant: float, num_qubits: int = 1) -> float:
+@deprecated(deadline='v2.0', fix='use cirq.qis.average_error')
+def average_error(decay_constant: float, num_qubits: int = 1) -> float:  # pragma: no cover
     """Calculates the average error from the depolarization decay constant.
 
     Args:
@@ -180,11 +187,13 @@ def average_error(decay_constant: float, num_qubits: int = 1) -> float:
     Returns:
         Calculated average error.
     """
-    N = 2**num_qubits
-    return (1 - decay_constant) * (1 - 1 / N)
+    return qis.average_error(decay_constant, num_qubits)
 
 
-def decoherence_pauli_error(t1_ns: float, tphi_ns: float, gate_time_ns: float) -> float:
+@deprecated(deadline='v2.0', fix='use cirq.qis.decoherence_pauli_error')
+def decoherence_pauli_error(
+    t1_ns: float, tphi_ns: float, gate_time_ns: float
+) -> float:  # pragma: no cover
     """The component of Pauli error caused by decoherence on a single qubit.
 
     Args:
@@ -195,11 +204,4 @@ def decoherence_pauli_error(t1_ns: float, tphi_ns: float, gate_time_ns: float) -
     Returns:
         Calculated Pauli error resulting from decoherence.
     """
-    gamma_2 = (1 / (2 * t1_ns)) + 1 / tphi_ns
-
-    exp1 = np.exp(-gate_time_ns / t1_ns)
-    exp2 = np.exp(-gate_time_ns * gamma_2)
-    px = 0.25 * (1 - exp1)
-    py = px
-    pz = 0.5 * (1 - exp2) - px
-    return px + py + pz
+    return qis.decoherence_pauli_error(t1_ns, tphi_ns, gate_time_ns)
diff --git a/cirq-core/cirq/devices/noise_utils_test.py b/cirq-core/cirq/devices/noise_utils_test.py
index df9833424ec..26aef22433f 100644
--- a/cirq-core/cirq/devices/noise_utils_test.py
+++ b/cirq-core/cirq/devices/noise_utils_test.py
@@ -12,20 +12,8 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import numpy as np
-import pytest
-
 import cirq
-from cirq.devices.noise_utils import (
-    OpIdentifier,
-    decay_constant_to_xeb_fidelity,
-    decay_constant_to_pauli_error,
-    pauli_error_to_decay_constant,
-    xeb_fidelity_to_decay_constant,
-    pauli_error_from_t1,
-    average_error,
-    decoherence_pauli_error,
-)
+from cirq.devices.noise_utils import OpIdentifier
 
 
 def test_op_identifier():
@@ -67,74 +55,3 @@ def test_op_id_instance():
     gate = cirq.SingleQubitCliffordGate.from_xz_map((cirq.X, False), (cirq.Z, False))
     op_id = OpIdentifier(gate, q0)
     cirq.testing.assert_equivalent_repr(op_id)
-
-
-@pytest.mark.parametrize(
-    'decay_constant,num_qubits,expected_output',
-    [(0.01, 1, 1 - (0.99 * 1 / 2)), (0.05, 2, 1 - (0.95 * 3 / 4))],
-)
-def test_decay_constant_to_xeb_fidelity(decay_constant, num_qubits, expected_output):
-    val = decay_constant_to_xeb_fidelity(decay_constant, num_qubits)
-    assert val == expected_output
-
-
-@pytest.mark.parametrize(
-    'decay_constant,num_qubits,expected_output',
-    [(0.01, 1, 0.99 * 3 / 4), (0.05, 2, 0.95 * 15 / 16)],
-)
-def test_decay_constant_to_pauli_error(decay_constant, num_qubits, expected_output):
-    val = decay_constant_to_pauli_error(decay_constant, num_qubits)
-    assert val == expected_output
-
-
-@pytest.mark.parametrize(
-    'pauli_error,num_qubits,expected_output',
-    [(0.01, 1, 1 - (0.01 / (3 / 4))), (0.05, 2, 1 - (0.05 / (15 / 16)))],
-)
-def test_pauli_error_to_decay_constant(pauli_error, num_qubits, expected_output):
-    val = pauli_error_to_decay_constant(pauli_error, num_qubits)
-    assert val == expected_output
-
-
-@pytest.mark.parametrize(
-    'xeb_fidelity,num_qubits,expected_output',
-    [(0.01, 1, 1 - 0.99 / (1 / 2)), (0.05, 2, 1 - 0.95 / (3 / 4))],
-)
-def test_xeb_fidelity_to_decay_constant(xeb_fidelity, num_qubits, expected_output):
-    val = xeb_fidelity_to_decay_constant(xeb_fidelity, num_qubits)
-    assert val == expected_output
-
-
-@pytest.mark.parametrize(
-    't,t1_ns,expected_output',
-    [
-        (20, 1e5, (1 - np.exp(-20 / 2e5)) / 2 + (1 - np.exp(-20 / 1e5)) / 4),
-        (4000, 1e4, (1 - np.exp(-4000 / 2e4)) / 2 + (1 - np.exp(-4000 / 1e4)) / 4),
-    ],
-)
-def test_pauli_error_from_t1(t, t1_ns, expected_output):
-    val = pauli_error_from_t1(t, t1_ns)
-    assert val == expected_output
-
-
-@pytest.mark.parametrize(
-    'decay_constant,num_qubits,expected_output', [(0.01, 1, 0.99 * 1 / 2), (0.05, 2, 0.95 * 3 / 4)]
-)
-def test_average_error(decay_constant, num_qubits, expected_output):
-    val = average_error(decay_constant, num_qubits)
-    assert val == expected_output
-
-
-@pytest.mark.parametrize(
-    'T1_ns,Tphi_ns,gate_time_ns', [(1e4, 2e4, 25), (1e5, 2e3, 25), (1e4, 2e4, 4000)]
-)
-def test_decoherence_pauli_error(T1_ns, Tphi_ns, gate_time_ns):
-    val = decoherence_pauli_error(T1_ns, Tphi_ns, gate_time_ns)
-    # Expected value is of the form:
-    #
-    #   (1/4) * [1 - e^(-t/T1)] + (1/2) * [1 - e^(-t/(2*T1) - t/Tphi]
-    #
-    expected_output = 0.25 * (1 - np.exp(-gate_time_ns / T1_ns)) + 0.5 * (
-        1 - np.exp(-gate_time_ns * ((1 / (2 * T1_ns)) + 1 / Tphi_ns))
-    )
-    assert val == expected_output
diff --git a/cirq-core/cirq/devices/superconducting_qubits_noise_properties.py b/cirq-core/cirq/devices/superconducting_qubits_noise_properties.py
index f1d1e2fdc3b..9a387162f48 100644
--- a/cirq-core/cirq/devices/superconducting_qubits_noise_properties.py
+++ b/cirq-core/cirq/devices/superconducting_qubits_noise_properties.py
@@ -20,7 +20,7 @@
 from functools import cached_property
 from typing import Dict, TYPE_CHECKING, List, Set, Type
 
-from cirq import ops, devices
+from cirq import ops, devices, qis
 from cirq.devices import noise_utils
 
 if TYPE_CHECKING:
@@ -129,7 +129,7 @@ def expected_gates(cls) -> Set[Type[ops.Gate]]:
     def _get_pauli_error(self, p_error: float, op_id: noise_utils.OpIdentifier):
         time_ns = float(self.gate_times_ns[op_id.gate_type])
         for q in op_id.qubits:
-            p_error -= noise_utils.decoherence_pauli_error(self.t1_ns[q], self.tphi_ns[q], time_ns)
+            p_error -= qis.decoherence_pauli_error(self.t1_ns[q], self.tphi_ns[q], time_ns)
         return p_error
 
     @cached_property
diff --git a/cirq-core/cirq/qis/__init__.py b/cirq-core/cirq/qis/__init__.py
index b2b8b0e4fcb..12913d6e8cc 100644
--- a/cirq-core/cirq/qis/__init__.py
+++ b/cirq-core/cirq/qis/__init__.py
@@ -50,3 +50,13 @@
     validate_qid_shape,
     validate_normalized_state_vector,
 )
+
+from cirq.qis.noise_utils import (
+    decay_constant_to_xeb_fidelity,
+    decay_constant_to_pauli_error,
+    pauli_error_to_decay_constant,
+    xeb_fidelity_to_decay_constant,
+    pauli_error_from_t1,
+    average_error,
+    decoherence_pauli_error,
+)
diff --git a/cirq-core/cirq/qis/noise_utils.py b/cirq-core/cirq/qis/noise_utils.py
new file mode 100644
index 00000000000..931f40de8ad
--- /dev/null
+++ b/cirq-core/cirq/qis/noise_utils.py
@@ -0,0 +1,123 @@
+# Copyright 2021 The Cirq Developers
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     https://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import numpy as np
+
+
+# TODO: expose all from top-level cirq?
+def decay_constant_to_xeb_fidelity(decay_constant: float, num_qubits: int = 2) -> float:
+    """Calculates the XEB fidelity from the depolarization decay constant.
+
+    Args:
+        decay_constant: Depolarization decay constant.
+        num_qubits: Number of qubits.
+
+    Returns:
+        Calculated XEB fidelity.
+    """
+    N = 2**num_qubits
+    return 1 - ((1 - decay_constant) * (1 - 1 / N))
+
+
+def decay_constant_to_pauli_error(decay_constant: float, num_qubits: int = 1) -> float:
+    """Calculates pauli error from the depolarization decay constant.
+
+    Args:
+        decay_constant: Depolarization decay constant.
+        num_qubits: Number of qubits.
+
+    Returns:
+        Calculated Pauli error.
+    """
+    N = 2**num_qubits
+    return (1 - decay_constant) * (1 - 1 / N / N)
+
+
+def pauli_error_to_decay_constant(pauli_error: float, num_qubits: int = 1) -> float:
+    """Calculates depolarization decay constant from pauli error.
+
+    Args:
+        pauli_error: The pauli error.
+        num_qubits: Number of qubits.
+
+    Returns:
+        Calculated depolarization decay constant.
+    """
+    N = 2**num_qubits
+    return 1 - (pauli_error / (1 - 1 / N / N))
+
+
+def xeb_fidelity_to_decay_constant(xeb_fidelity: float, num_qubits: int = 2) -> float:
+    """Calculates the depolarization decay constant from XEB fidelity.
+
+    Args:
+        xeb_fidelity: The XEB fidelity.
+        num_qubits: Number of qubits.
+
+    Returns:
+        Calculated depolarization decay constant.
+    """
+    N = 2**num_qubits
+    return 1 - (1 - xeb_fidelity) / (1 - 1 / N)
+
+
+def pauli_error_from_t1(t_ns: float, t1_ns: float) -> float:
+    """Calculates the pauli error from T1 decay constant.
+
+    This computes error for a specific duration, `t`.
+
+    Args:
+        t_ns: The duration of the gate in ns.
+        t1_ns: The T1 decay constant in ns.
+
+    Returns:
+        Calculated Pauli error resulting from T1 decay.
+    """
+    t2 = 2 * t1_ns
+    return (1 - np.exp(-t_ns / t2)) / 2 + (1 - np.exp(-t_ns / t1_ns)) / 4
+
+
+def average_error(decay_constant: float, num_qubits: int = 1) -> float:
+    """Calculates the average error from the depolarization decay constant.
+
+    Args:
+        decay_constant: Depolarization decay constant.
+        num_qubits: Number of qubits.
+
+    Returns:
+        Calculated average error.
+    """
+    N = 2**num_qubits
+    return (1 - decay_constant) * (1 - 1 / N)
+
+
+def decoherence_pauli_error(t1_ns: float, tphi_ns: float, gate_time_ns: float) -> float:
+    """The component of Pauli error caused by decoherence on a single qubit.
+
+    Args:
+        t1_ns: T1 time in nanoseconds.
+        tphi_ns: Tphi time in nanoseconds.
+        gate_time_ns: Duration in nanoseconds of the gate affected by this error.
+
+    Returns:
+        Calculated Pauli error resulting from decoherence.
+    """
+    gamma_2 = (1 / (2 * t1_ns)) + 1 / tphi_ns
+
+    exp1 = np.exp(-gate_time_ns / t1_ns)
+    exp2 = np.exp(-gate_time_ns * gamma_2)
+    px = 0.25 * (1 - exp1)
+    py = px
+    pz = 0.5 * (1 - exp2) - px
+    return px + py + pz
diff --git a/cirq-core/cirq/qis/noise_utils_test.py b/cirq-core/cirq/qis/noise_utils_test.py
new file mode 100644
index 00000000000..d7ed1bd5f7d
--- /dev/null
+++ b/cirq-core/cirq/qis/noise_utils_test.py
@@ -0,0 +1,97 @@
+# Copyright 2021 The Cirq Developers
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     https://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import numpy as np
+import pytest
+
+from cirq.qis.noise_utils import (
+    decay_constant_to_xeb_fidelity,
+    decay_constant_to_pauli_error,
+    pauli_error_to_decay_constant,
+    xeb_fidelity_to_decay_constant,
+    pauli_error_from_t1,
+    average_error,
+    decoherence_pauli_error,
+)
+
+
+@pytest.mark.parametrize(
+    'decay_constant,num_qubits,expected_output',
+    [(0.01, 1, 1 - (0.99 * 1 / 2)), (0.05, 2, 1 - (0.95 * 3 / 4))],
+)
+def test_decay_constant_to_xeb_fidelity(decay_constant, num_qubits, expected_output):
+    val = decay_constant_to_xeb_fidelity(decay_constant, num_qubits)
+    assert val == expected_output
+
+
+@pytest.mark.parametrize(
+    'decay_constant,num_qubits,expected_output',
+    [(0.01, 1, 0.99 * 3 / 4), (0.05, 2, 0.95 * 15 / 16)],
+)
+def test_decay_constant_to_pauli_error(decay_constant, num_qubits, expected_output):
+    val = decay_constant_to_pauli_error(decay_constant, num_qubits)
+    assert val == expected_output
+
+
+@pytest.mark.parametrize(
+    'pauli_error,num_qubits,expected_output',
+    [(0.01, 1, 1 - (0.01 / (3 / 4))), (0.05, 2, 1 - (0.05 / (15 / 16)))],
+)
+def test_pauli_error_to_decay_constant(pauli_error, num_qubits, expected_output):
+    val = pauli_error_to_decay_constant(pauli_error, num_qubits)
+    assert val == expected_output
+
+
+@pytest.mark.parametrize(
+    'xeb_fidelity,num_qubits,expected_output',
+    [(0.01, 1, 1 - 0.99 / (1 / 2)), (0.05, 2, 1 - 0.95 / (3 / 4))],
+)
+def test_xeb_fidelity_to_decay_constant(xeb_fidelity, num_qubits, expected_output):
+    val = xeb_fidelity_to_decay_constant(xeb_fidelity, num_qubits)
+    assert val == expected_output
+
+
+@pytest.mark.parametrize(
+    't,t1_ns,expected_output',
+    [
+        (20, 1e5, (1 - np.exp(-20 / 2e5)) / 2 + (1 - np.exp(-20 / 1e5)) / 4),
+        (4000, 1e4, (1 - np.exp(-4000 / 2e4)) / 2 + (1 - np.exp(-4000 / 1e4)) / 4),
+    ],
+)
+def test_pauli_error_from_t1(t, t1_ns, expected_output):
+    val = pauli_error_from_t1(t, t1_ns)
+    assert val == expected_output
+
+
+@pytest.mark.parametrize(
+    'decay_constant,num_qubits,expected_output', [(0.01, 1, 0.99 * 1 / 2), (0.05, 2, 0.95 * 3 / 4)]
+)
+def test_average_error(decay_constant, num_qubits, expected_output):
+    val = average_error(decay_constant, num_qubits)
+    assert val == expected_output
+
+
+@pytest.mark.parametrize(
+    'T1_ns,Tphi_ns,gate_time_ns', [(1e4, 2e4, 25), (1e5, 2e3, 25), (1e4, 2e4, 4000)]
+)
+def test_decoherence_pauli_error(T1_ns, Tphi_ns, gate_time_ns):
+    val = decoherence_pauli_error(T1_ns, Tphi_ns, gate_time_ns)
+    # Expected value is of the form:
+    #
+    #   (1/4) * [1 - e^(-t/T1)] + (1/2) * [1 - e^(-t/(2*T1) - t/Tphi]
+    #
+    expected_output = 0.25 * (1 - np.exp(-gate_time_ns / T1_ns)) + 0.5 * (
+        1 - np.exp(-gate_time_ns * ((1 / (2 * T1_ns)) + 1 / Tphi_ns))
+    )
+    assert val == expected_output