diff --git a/quaddtype/numpy_quaddtype/src/umath/unary_ops.cpp b/quaddtype/numpy_quaddtype/src/umath/unary_ops.cpp
index d60591e8..b8a82aee 100644
--- a/quaddtype/numpy_quaddtype/src/umath/unary_ops.cpp
+++ b/quaddtype/numpy_quaddtype/src/umath/unary_ops.cpp
@@ -156,6 +156,11 @@ init_quad_unary_ops(PyObject *numpy)
     if (create_quad_unary_ufunc<quad_absolute, ld_absolute>(numpy, "absolute") < 0) {
         return -1;
     }
+    // fabs is simialr to absolute, just not handles complex values (we neither)
+    // registering the absolute here
+    if (create_quad_unary_ufunc<quad_absolute, ld_absolute>(numpy, "fabs") < 0) {
+        return -1;
+    }
     if (create_quad_unary_ufunc<quad_sign, ld_sign>(numpy, "sign") < 0) {
         return -1;
     }
diff --git a/quaddtype/release_tracker.md b/quaddtype/release_tracker.md
index 74c75315..4e3e9068 100644
--- a/quaddtype/release_tracker.md
+++ b/quaddtype/release_tracker.md
@@ -23,7 +23,7 @@
 | fmod          | ✅    | ✅                                                                   |
 | divmod        |       |                                                                      |
 | absolute      | ✅    | ✅                                                                   |
-| fabs          |       |                                                                      |
+| fabs          | ✅    | ✅                                                                   |
 | rint          | ✅    | ✅                                                                   |
 | sign          | ✅    | ✅                                                                   |
 | heaviside     |       |                                                                      |
diff --git a/quaddtype/tests/test_quaddtype.py b/quaddtype/tests/test_quaddtype.py
index 3c765b5b..ca500f49 100644
--- a/quaddtype/tests/test_quaddtype.py
+++ b/quaddtype/tests/test_quaddtype.py
@@ -1349,4 +1349,42 @@ def test_float_power_array():
 
     assert result.dtype.name == "QuadPrecDType128"
     for i in range(len(result)):
-        np.testing.assert_allclose(float(result[i]), expected[i], rtol=1e-13)
\ No newline at end of file
+        np.testing.assert_allclose(float(result[i]), expected[i], rtol=1e-13)
+
+
+@pytest.mark.parametrize("val", [
+    # Positive values
+    "3.0", "12.5", "100.0", "1e100", "0.0",
+    # Negative values
+    "-3.0", "-12.5", "-100.0", "-1e100", "-0.0",
+    # Special values
+    "inf", "-inf", "nan", "-nan",
+    # Small values
+    "1e-100", "-1e-100"
+])
+def test_fabs(val):
+    """
+    Test np.fabs ufunc for QuadPrecision dtype.
+    fabs computes absolute values (positive magnitude) for floating-point numbers.
+    It should behave identically to np.absolute for real (non-complex) types.
+    """
+    quad_val = QuadPrecision(val)
+    float_val = float(val)
+
+    quad_result = np.fabs(quad_val)
+    float_result = np.fabs(float_val)
+
+    # Test with both scalar and array
+    quad_arr = np.array([quad_val], dtype=QuadPrecDType())
+    quad_arr_result = np.fabs(quad_arr)
+
+    # Check scalar result
+    np.testing.assert_array_equal(np.array(quad_result).astype(float), float_result)
+
+    # Check array result
+    np.testing.assert_array_equal(quad_arr_result.astype(float)[0], float_result)
+
+    # For zero results, check sign (should always be positive after fabs)
+    if float_result == 0.0:
+        assert not np.signbit(quad_result), f"fabs({val}) should not have negative sign"
+        assert not np.signbit(quad_arr_result[0]), f"fabs({val}) should not have negative sign"