diff --git a/src/utility/include/mp-units/math.h b/src/utility/include/mp-units/math.h
index 7bdf71893..29254672e 100644
--- a/src/utility/include/mp-units/math.h
+++ b/src/utility/include/mp-units/math.h
@@ -271,6 +271,19 @@ template<Unit auto To, auto R, typename Rep>
   }
 }
 
+/**
+ * @brief Computes the inverse of a quantity in a provided unit
+ */
+template<Unit auto To, auto R, typename Rep>
+[[nodiscard]] constexpr QuantityOf<dimensionless / get_quantity_spec(R)> auto inverse(const quantity<R, Rep>& q)
+  requires requires {
+    quantity_values<Rep>::one();
+    value_cast<To>(1 / q);
+  }
+{
+  return (quantity_values<Rep>::one() * one).force_in(To * q.unit) / q;
+}
+
 /**
  * @brief Computes the square root of the sum of the squares of x and y,
  *        without undue overflow or underflow at intermediate stages of the computation
diff --git a/test/unit_test/static/math_test.cpp b/test/unit_test/static/math_test.cpp
index a8b23053a..980f68e08 100644
--- a/test/unit_test/static/math_test.cpp
+++ b/test/unit_test/static/math_test.cpp
@@ -26,8 +26,6 @@
 #include <mp-units/systems/si/unit_symbols.h>
 #include <optional>
 
-#if __cpp_lib_constexpr_cmath || MP_UNITS_COMP_GCC
-
 namespace {
 
 using namespace mp_units;
@@ -40,6 +38,8 @@ template<typename T1, typename T2, typename... Ts>
   return is_same_v<T1, T2> && v1 == v2 && (... && (v1 == vs));
 }
 
+#if __cpp_lib_constexpr_cmath || MP_UNITS_COMP_GCC
+
 static_assert(compare(pow<0>(2 * m), 1 * one));
 static_assert(compare(pow<1>(2 * m), 2 * m));
 static_assert(compare(pow<2>(2 * m), 4 * pow<2>(m), 4 * m2));
@@ -201,6 +201,20 @@ static_assert(compare(round<si::second>(-1499. * isq::time[ms]), -1. * isq::time
 static_assert(compare(round<si::second>(-1500. * isq::time[ms]), -2. * isq::time[s]));
 static_assert(compare(round<si::second>(-1999. * isq::time[ms]), -2. * isq::time[s]));
 
-}  // namespace
-
 #endif
+
+// non-truncating
+static_assert(compare(inverse<us>(250 * Hz), 4000 * us));
+static_assert(compare(inverse<us>(250 * kHz), 4 * us));
+static_assert(compare(inverse<ks>(250 * uHz), 4 * ks));
+
+// truncating
+static_assert(compare(inverse<s>(1 * kHz), 0 * s));
+
+// floating-point representation does not truncate
+static_assert(compare(inverse<s>(1. * kHz), 0.001 * s));
+
+// check if constraints work properly for a derived unit of a narrowed kind
+static_assert(compare(inverse<Hz>(1 * s), 1 * Hz));
+
+}  // namespace