diff --git a/Cargo.toml b/Cargo.toml
index 37df7c5..98a8483 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -20,6 +20,13 @@ publish = false
 [package.metadata.docs.rs]
 rustdoc-args = [ "--html-in-header", "katex-header.html" ]
 
+[dev-dependencies]
+criterion = "0.3"
+
+[[bench]]
+name = "arithmetic"
+harness = false
+
 [dependencies]
 subtle = "2.2.1"
 crossbeam-utils = "0.7"
diff --git a/benches/arithmetic.rs b/benches/arithmetic.rs
new file mode 100644
index 0000000..49d965b
--- /dev/null
+++ b/benches/arithmetic.rs
@@ -0,0 +1,34 @@
+#[macro_use]
+extern crate criterion;
+
+extern crate halo2;
+use crate::arithmetic::{small_multiexp, EqAffine, Field, Fp, Fq};
+use crate::poly::commitment::Params;
+use crate::transcript::DummyHash;
+use halo2::*;
+
+use criterion::{black_box, Criterion};
+
+fn criterion_benchmark(c: &mut Criterion) {
+    // small multiexp
+    {
+        let params: Params<EqAffine> = Params::new::<DummyHash<Fq>>(5);
+        let g = &mut params.get_g();
+        let len = g.len() / 2;
+        let (g_lo, g_hi) = g.split_at_mut(len);
+
+        let coeff_1 = Fp::random();
+        let coeff_2 = Fp::random();
+
+        c.bench_function("double-and-add", |b| {
+            b.iter(|| {
+                for (g_lo, g_hi) in g_lo.iter().zip(g_hi.iter()) {
+                    small_multiexp(&[black_box(coeff_1), black_box(coeff_2)], &[*g_lo, *g_hi]);
+                }
+            })
+        });
+    }
+}
+
+criterion_group!(benches, criterion_benchmark);
+criterion_main!(benches);
diff --git a/src/arithmetic.rs b/src/arithmetic.rs
index 312cfe5..a7193a8 100644
--- a/src/arithmetic.rs
+++ b/src/arithmetic.rs
@@ -180,6 +180,30 @@ fn multiexp_serial<C: CurveAffine>(coeffs: &[C::Scalar], bases: &[C], acc: &mut
     }
 }
 
+/// Performs a small multi-exponentiation operation.
+/// Uses the double-and-add algorithm with doublings shared across points.
+pub fn small_multiexp<C: CurveAffine>(coeffs: &[C::Scalar], bases: &[C]) -> C::Projective {
+    let coeffs: Vec<[u8; 32]> = coeffs.iter().map(|a| a.to_bytes()).collect();
+    let mut acc = C::Projective::zero();
+
+    // for byte idx
+    for byte_idx in (0..32).rev() {
+        // for bit idx
+        for bit_idx in (0..8).rev() {
+            acc = acc.double();
+            // for each coeff
+            for coeff_idx in 0..coeffs.len() {
+                let byte = coeffs[coeff_idx][byte_idx];
+                if ((byte >> bit_idx) & 1) != 0 {
+                    acc += bases[coeff_idx];
+                }
+            }
+        }
+    }
+
+    acc
+}
+
 /// Performs a multi-exponentiation operation.
 ///
 /// This function will panic if coeffs and bases have a different length.
diff --git a/src/poly/commitment.rs b/src/poly/commitment.rs
index a1a9f60..2eaa9ef 100644
--- a/src/poly/commitment.rs
+++ b/src/poly/commitment.rs
@@ -275,6 +275,11 @@ impl<C: CurveAffine> Params<C> {
             other_bases,
         }
     }
+
+    /// Getter for g generators
+    pub fn get_g(&self) -> Vec<C> {
+        self.g.clone()
+    }
 }
 
 /// A guard returned by the verifier
diff --git a/src/poly/commitment/prover.rs b/src/poly/commitment/prover.rs
index 3acddc2..c66fe82 100644
--- a/src/poly/commitment/prover.rs
+++ b/src/poly/commitment/prover.rs
@@ -1,8 +1,8 @@
 use super::super::{Coeff, Polynomial};
 use super::{Blind, OpeningProof, Params};
 use crate::arithmetic::{
-    best_multiexp, compute_inner_product, get_challenge_scalar, parallelize, Challenge, Curve,
-    CurveAffine, Field,
+    best_multiexp, compute_inner_product, get_challenge_scalar, parallelize, small_multiexp,
+    Challenge, Curve, CurveAffine, Field,
 };
 use crate::transcript::Hasher;
 
@@ -226,8 +226,7 @@ fn parallel_generator_collapse<C: CurveAffine>(
         let g_hi = &g_hi[start..];
         let mut tmp = Vec::with_capacity(g_lo.len());
         for (g_lo, g_hi) in g_lo.iter().zip(g_hi.iter()) {
-            // TODO: could use multiexp
-            tmp.push(((*g_lo) * challenge_inv) + &((*g_hi) * challenge));
+            tmp.push(small_multiexp(&[challenge_inv, challenge], &[*g_lo, *g_hi]));
         }
         C::Projective::batch_to_affine(&tmp, g_lo);
     });