diff --git a/Cargo.lock b/Cargo.lock
index 86e8dc0650..5c4c0d7c5a 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -2179,6 +2179,7 @@ dependencies = [
  "ark-serialize",
  "blake2",
  "colored",
+ "criterion",
  "groupmap",
  "itertools 0.12.1",
  "mina-curves",
diff --git a/poly-commitment/Cargo.toml b/poly-commitment/Cargo.toml
index 680c50cc99..5496e686ae 100644
--- a/poly-commitment/Cargo.toml
+++ b/poly-commitment/Cargo.toml
@@ -35,9 +35,14 @@ ocaml = { workspace = true, optional = true }
 ocaml-gen = { workspace = true, optional = true }
 
 [dev-dependencies]
+criterion.workspace = true
 colored.workspace = true
 rand_chacha.workspace = true
 ark-bn254.workspace = true
 
 [features]
 ocaml_types = ["ocaml", "ocaml-gen"]
+
+[[bench]]
+name = "poly_comm"
+harness = false
\ No newline at end of file
diff --git a/poly-commitment/benches/poly_comm.rs b/poly-commitment/benches/poly_comm.rs
new file mode 100644
index 0000000000..5a961d6c7c
--- /dev/null
+++ b/poly-commitment/benches/poly_comm.rs
@@ -0,0 +1,112 @@
+use ark_ec::{
+    short_weierstrass::{Affine, SWCurveConfig},
+    AffineRepr,
+};
+use ark_ff::{PrimeField, UniformRand};
+use criterion::{black_box, criterion_group, criterion_main, BatchSize, BenchmarkId, Criterion};
+use mina_curves::pasta::{Pallas, Vesta};
+use poly_commitment::PolyComm;
+use rand_core::{CryptoRng, RngCore};
+
+fn helper_generate_random_elliptic_curve_point<RNG, P>(rng: &mut RNG) -> Affine<P>
+where
+    P::BaseField: PrimeField,
+    RNG: RngCore + CryptoRng,
+    P: SWCurveConfig,
+{
+    let p1_x = P::BaseField::rand(rng);
+    let mut p1: Option<Affine<P>> = Affine::<P>::get_point_from_x_unchecked(p1_x, false);
+    while p1.is_none() {
+        let p1_x = P::BaseField::rand(rng);
+        p1 = Affine::<P>::get_point_from_x_unchecked(p1_x, false);
+    }
+    let p1: Affine<P> = p1.unwrap().mul_by_cofactor_to_group().into();
+    p1
+}
+
+fn generate_poly_comm_pallas<RNG>(rng: &mut RNG, n: usize) -> PolyComm<Pallas>
+where
+    RNG: RngCore + CryptoRng,
+{
+    let elems: Vec<Pallas> = (0..n)
+        .map(|_| helper_generate_random_elliptic_curve_point(rng))
+        .collect();
+    PolyComm::new(elems)
+}
+
+fn generate_poly_comm_vesta<RNG>(rng: &mut RNG, n: usize) -> PolyComm<Vesta>
+where
+    RNG: RngCore + CryptoRng,
+{
+    let elems: Vec<Vesta> = (0..n)
+        .map(|_| helper_generate_random_elliptic_curve_point(rng))
+        .collect();
+    PolyComm::new(elems)
+}
+
+fn benchmark_polycomm_add(c: &mut Criterion) {
+    let mut group = c.benchmark_group("PolyComm Add");
+    let mut rng = o1_utils::tests::make_test_rng(None);
+
+    for n in [16, 32, 64, 128, 256, 1 << 10, 1 << 14, 1 << 15].into_iter() {
+        let poly1: PolyComm<Pallas> = generate_poly_comm_pallas(&mut rng, n);
+        let poly2: PolyComm<Pallas> = generate_poly_comm_pallas(&mut rng, n);
+        group.bench_with_input(BenchmarkId::new("PolyComm Add Pallas", n), &n, |b, _| {
+            b.iter_batched(
+                || (poly1.clone(), poly2.clone()),
+                |(poly1, poly2)| {
+                    black_box(&poly1 + &poly2);
+                },
+                BatchSize::SmallInput,
+            )
+        });
+
+        let poly1: PolyComm<Vesta> = generate_poly_comm_vesta(&mut rng, n);
+        let poly2: PolyComm<Vesta> = generate_poly_comm_vesta(&mut rng, n);
+        group.bench_with_input(BenchmarkId::new("PolyComm Add Vesta", n), &n, |b, _| {
+            b.iter_batched(
+                || (poly1.clone(), poly2.clone()),
+                |(poly1, poly2)| {
+                    black_box(&poly1 + &poly2);
+                },
+                BatchSize::SmallInput,
+            )
+        });
+    }
+    group.finish();
+}
+
+fn benchmark_polycomm_sub(c: &mut Criterion) {
+    let mut group = c.benchmark_group("PolyComm Sub");
+    let mut rng = o1_utils::tests::make_test_rng(None);
+
+    for n in [16, 32, 64, 128, 256, 1 << 10, 1 << 14, 1 << 15].into_iter() {
+        let poly1: PolyComm<Pallas> = generate_poly_comm_pallas(&mut rng, n);
+        let poly2: PolyComm<Pallas> = generate_poly_comm_pallas(&mut rng, n);
+        group.bench_with_input(BenchmarkId::new("PolyComm Sub Pallas", n), &n, |b, _| {
+            b.iter_batched(
+                || (poly1.clone(), poly2.clone()),
+                |(poly1, poly2)| {
+                    black_box(&poly1 - &poly2);
+                },
+                BatchSize::SmallInput,
+            )
+        });
+
+        let poly1: PolyComm<Vesta> = generate_poly_comm_vesta(&mut rng, n);
+        let poly2: PolyComm<Vesta> = generate_poly_comm_vesta(&mut rng, n);
+        group.bench_with_input(BenchmarkId::new("PolyComm Sub Vesta", n), &n, |b, _| {
+            b.iter_batched(
+                || (poly1.clone(), poly2.clone()),
+                |(poly1, poly2)| {
+                    black_box(&poly1 - &poly2);
+                },
+                BatchSize::SmallInput,
+            )
+        });
+    }
+    group.finish();
+}
+
+criterion_group!(benches, benchmark_polycomm_add, benchmark_polycomm_sub);
+criterion_main!(benches);