diff --git a/src/uu/sort/Cargo.toml b/src/uu/sort/Cargo.toml
index 09763a4f9..d789d68b4 100644
--- a/src/uu/sort/Cargo.toml
+++ b/src/uu/sort/Cargo.toml
@@ -81,6 +81,10 @@ harness = false
 name = "sort_locale_utf8_bench"
 harness = false
 
+[[bench]]
+name = "sort_bench_merge"
+harness = false
+
 [[bench]]
 name = "sort_locale_de_bench"
 harness = false
diff --git a/src/uu/sort/benches/sort_bench_merge.rs b/src/uu/sort/benches/sort_bench_merge.rs
new file mode 100644
index 000000000..9189c9778
--- /dev/null
+++ b/src/uu/sort/benches/sort_bench_merge.rs
@@ -0,0 +1,70 @@
+// This file is part of the uutils coreutils package.
+//
+// For the full copyright and license information, please view the LICENSE
+// file that was distributed with this source code.
+
+use divan::{Bencher, black_box};
+use tempfile::NamedTempFile;
+use uu_sort::uumain;
+use uucore::benchmark::{run_util_function, setup_test_file, text_data};
+
+/// Benchmark merging pre-sorted files (`sort -m`).
+///
+/// Mirrors the "Merging" section from `BENCHMARKING.md`:
+/// 1. Generate a shuffled wordlist (using the existing ASCII helper).
+/// 2. Split it into slices.
+/// 3. Sort each slice individually.
+/// 4. Benchmark merging them back together with `sort -m`.
+#[divan::bench]
+fn merge_pre_sorted_files(bencher: Bencher) {
+    const TOTAL_LINES: usize = 500_000;
+    const NUM_FILES: usize = 8;
+
+    // 1. Generate data (mimicking a shuffled wordlist)
+    let data = text_data::generate_ascii_data(TOTAL_LINES);
+
+    // 2. Split into chunks and 3. sort each chunk individually
+    let lines: Vec<&[u8]> = data
+        .split(|&b| b == b'\n')
+        .filter(|l| !l.is_empty())
+        .collect();
+    let lines_per_file = lines.len() / NUM_FILES;
+
+    let mut file_paths = Vec::with_capacity(NUM_FILES);
+    for i in 0..NUM_FILES {
+        let start = i * lines_per_file;
+        let end = if i == NUM_FILES - 1 {
+            lines.len()
+        } else {
+            (i + 1) * lines_per_file
+        };
+
+        let mut chunk_lines = lines[start..end].to_vec();
+        chunk_lines.sort();
+
+        let mut chunk_data = Vec::new();
+        for line in chunk_lines {
+            chunk_data.extend_from_slice(line);
+            chunk_data.push(b'\n');
+        }
+
+        let path = setup_test_file(&chunk_data);
+        file_paths.push(path);
+    }
+
+    let output_file = NamedTempFile::new().unwrap();
+    let output_path = output_file.path().to_str().unwrap();
+
+    let file_args: Vec<&str> = file_paths.iter().map(|p| p.to_str().unwrap()).collect();
+
+    // 4. Benchmark the merge step
+    bencher.bench(|| {
+        let mut args = vec!["-m", "-o", output_path];
+        args.extend(&file_args);
+        black_box(run_util_function(uumain, &args));
+    });
+}
+
+fn main() {
+    divan::main();
+}