diff --git a/libcxx/test/benchmarks/algorithms/sorting/nth_element.bench.cpp b/libcxx/test/benchmarks/algorithms/sorting/nth_element.bench.cpp
new file mode 100644
index 0000000000000..4c3e8115def30
--- /dev/null
+++ b/libcxx/test/benchmarks/algorithms/sorting/nth_element.bench.cpp
@@ -0,0 +1,91 @@
+//===----------------------------------------------------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+// UNSUPPORTED: c++03, c++11, c++14, c++17
+
+#include <algorithm>
+#include <array>
+#include <cstddef>
+#include <deque>
+#include <string>
+#include <vector>
+
+#include "benchmark/benchmark.h"
+#include "common.h"
+
+int main(int argc, char** argv) {
+  auto std_nth_element = [](auto first, auto nth, auto last) { return std::nth_element(first, nth, last); };
+
+  // Benchmark std::nth_element on various types of data. We always partition
+  // around the middle element of the full range. We don't benchmark
+  // std::ranges::nth_element separately because it forwards to the same core.
+  //
+  // We run this in batches because nth_element mutates the range in place, so we
+  // restore the original data after each batch.
+  //
+  // We intentionally don't benchmark the predicated version of the algorithm
+  // because that makes the benchmark run too slowly.
+  {
+    auto bm = []<class Container>(std::string name, auto nth_element, auto generate_data) {
+      benchmark::RegisterBenchmark(
+          name,
+          [nth_element, generate_data](auto& st) {
+            std::size_t const size          = st.range(0);
+            constexpr std::size_t BatchSize = 32;
+            using ValueType                 = typename Container::value_type;
+            std::vector<ValueType> data     = generate_data(size);
+            std::array<Container, BatchSize> c;
+            std::fill_n(c.begin(), BatchSize, Container(data.begin(), data.end()));
+
+            std::size_t const half = size / 2;
+            while (st.KeepRunningBatch(BatchSize)) {
+              for (std::size_t i = 0; i != BatchSize; ++i) {
+                benchmark::DoNotOptimize(c[i]);
+                nth_element(c[i].begin(), c[i].begin() + half, c[i].end());
+                benchmark::DoNotOptimize(c[i]);
+              }
+
+              st.PauseTiming();
+              for (std::size_t i = 0; i != BatchSize; ++i) {
+                std::copy(data.begin(), data.end(), c[i].begin());
+              }
+              st.ResumeTiming();
+            }
+          })
+          ->Arg(8)
+          ->Arg(1024)
+          ->Arg(8192);
+    };
+
+    auto register_bm = [&](auto generate, std::string variant) {
+      auto gen2 = [generate](auto size) {
+        std::vector<int> data = generate(size);
+        std::vector<support::NonIntegral> real_data(data.begin(), data.end());
+        return real_data;
+      };
+      auto name = [variant](std::string op) { return op + " (" + variant + ")"; };
+      bm.operator()<std::vector<int>>(name("std::nth_element(vector<int>)"), std_nth_element, generate);
+      bm.operator()<std::vector<support::NonIntegral>>(
+          name("std::nth_element(vector<NonIntegral>)"), std_nth_element, gen2);
+      bm.operator()<std::deque<int>>(name("std::nth_element(deque<int>)"), std_nth_element, generate);
+    };
+
+    register_bm(support::quicksort_adversarial_data<int>, "qsort adversarial");
+    register_bm(support::ascending_sorted_data<int>, "ascending");
+    register_bm(support::descending_sorted_data<int>, "descending");
+    register_bm(support::pipe_organ_data<int>, "pipe-organ");
+    register_bm(support::heap_data<int>, "heap");
+    register_bm(support::shuffled_data<int>, "shuffled");
+    register_bm(support::single_element_data<int>, "repeated");
+  }
+
+  benchmark::Initialize(&argc, argv);
+  benchmark::RunSpecifiedBenchmarks();
+  benchmark::Shutdown();
+  return 0;
+}