diff --git a/RELEASE.md b/RELEASE.md
index 6674e437552923..a8bbc7064bfda4 100644
--- a/RELEASE.md
+++ b/RELEASE.md
@@ -57,6 +57,8 @@
 * tf.image.decode_jpeg by default uses the faster DCT method, sacrificing
   a little fidelity for improved speed. One can revert to the old
   behavior by specifying the attribute dct_method='INTEGER_ACCURATE'.
+* `tf.complex_abs` has been removed from the Python interface. `tf.abs`
+  supports complex tensors and should be used instead.
 
 # Release 0.12.0
 
diff --git a/tensorflow/core/ops/math_ops.cc b/tensorflow/core/ops/math_ops.cc
index 04065849539e52..81cacffe96421c 100644
--- a/tensorflow/core/ops/math_ops.cc
+++ b/tensorflow/core/ops/math_ops.cc
@@ -175,13 +175,6 @@ Given a tensor `x` of complex numbers, this operation returns a tensor of type
 `float` or `double` that is the absolute value of each element in `x`. All
 elements in `x` must be complex numbers of the form \\(a + bj\\). The absolute
 value is computed as \\( \sqrt{a^2 + b^2}\\).
-
-For example:
-
-```
-# tensor 'x' is [[-2.25 + 4.75j], [-3.25 + 5.75j]]
-tf.complex_abs(x) ==> [5.25594902, 6.60492229]
-```
 )doc");
 
 // Declares cwise unary operations signature: 't -> 't
diff --git a/tensorflow/python/kernel_tests/cwise_ops_test.py b/tensorflow/python/kernel_tests/cwise_ops_test.py
index 79ec1ab62790ed..a828273314cd68 100644
--- a/tensorflow/python/kernel_tests/cwise_ops_test.py
+++ b/tensorflow/python/kernel_tests/cwise_ops_test.py
@@ -372,7 +372,7 @@ def testComplex64Basic(self):
     x = np.complex(1, 1) * np.arange(-3, 3).reshape(1, 3,
                                                     2).astype(np.complex64)
     y = x + 0.5  # no zeros
-    self._compareCpu(x, np.abs, math_ops.complex_abs)
+    self._compareCpu(x, np.abs, math_ops.abs)
     self._compareCpu(x, np.abs, _ABS)
     self._compareCpu(x, np.negative, math_ops.neg)
     self._compareCpu(x, np.negative, _NEG)
@@ -1932,7 +1932,7 @@ def _compareBroadcastGradient(self, x):
     epsilon = 1e-3
     with self.test_session():
       for args in [(x_, 0.), (0., x_)]:
-        z = math_ops.reduce_sum(math_ops.complex_abs(math_ops.complex(*args)))
+        z = math_ops.reduce_sum(math_ops.abs(math_ops.complex(*args)))
         jacob_t, jacob_n = gradient_checker.compute_gradient(
             x_, list(x.shape), z, [1], x_init_value=x, delta=epsilon)
         self.assertAllClose(jacob_t, jacob_n, rtol=epsilon, atol=epsilon)
diff --git a/tensorflow/python/ops/math_grad_test.py b/tensorflow/python/ops/math_grad_test.py
index 6678ae4e26d1b9..8b0bc6b4f18cf1 100644
--- a/tensorflow/python/ops/math_grad_test.py
+++ b/tensorflow/python/ops/math_grad_test.py
@@ -77,10 +77,7 @@ def _testGrad(self, shape, dtype=None, max_error=None, bias=None, sigma=None):
               shape, bias=bias), dtype=dtype)
 
     with self.test_session(use_gpu=True):
-      if dtype in (dtypes.complex64, dtypes.complex128):
-        output = math_ops.complex_abs(value)
-      else:
-        output = math_ops.abs(value)
+      output = math_ops.abs(value)
       error = gradient_checker.compute_gradient_error(
           value, shape, output, output.get_shape().as_list())
     self.assertLess(error, max_error)
diff --git a/tensorflow/python/ops/math_ops.py b/tensorflow/python/ops/math_ops.py
index 0f7727ae686cc2..20b9e438a1f265 100644
--- a/tensorflow/python/ops/math_ops.py
+++ b/tensorflow/python/ops/math_ops.py
@@ -125,7 +125,6 @@
 functions to your graph.
 
 @@complex
-@@complex_abs
 @@conj
 @@imag
 @@real
@@ -439,31 +438,6 @@ def erf(x, name=None):
       return gen_math_ops.erf(x, name=name)
 
 
-def complex_abs(x, name=None):
-  r"""Computes the complex absolute value of a tensor.
-
-  Given a tensor `x` of complex numbers, this operation returns a tensor of type
-  `float32` or `float64` that is the absolute value of each element in `x`. All
-  elements in `x` must be complex numbers of the form \\(a + bj\\). The
-  absolute value is computed as \\( \sqrt{a^2 + b^2}\\).
-
-  For example:
-
-  ```
-  # tensor 'x' is [[-2.25 + 4.75j], [-3.25 + 5.75j]]
-  tf.complex_abs(x) ==> [5.25594902, 6.60492229]
-  ```
-
-  Args:
-    x: A `Tensor` of type `complex64` or `complex128`.
-    name: A name for the operation (optional).
-
-  Returns:
-    A `Tensor` of type `float32` or `float64`.
-  """
-  return gen_math_ops._complex_abs(x, Tout=x.dtype.real_dtype, name=name)
-
-
 def scalar_mul(scalar, x):
   """Multiplies a scalar times a `Tensor` or `IndexedSlices` object.