Quantum computers compared
15 machines, 9 vendors
15 quantum computers, each with the figures its vendor publishes and the rates a run on it is sold for. The fourth column counts how many of the 9 circuit families measured here each could run at 20 qubits and depth 3.
Every quantum computer
What it is and where a run is sold
| Quantum computer | What it is | Two-qubit gate error | Circuit families it can run at 20 qubits | Where a run is sold |
|---|---|---|---|---|
| What it is 36 trapped ions; all-to-all | Two-qubit gate error 0.96% | Circuit families it can run at 20 qubits 2 of 9 | Where a run is sold | |
| What it is 36 trapped ions; all-to-all; on-prem | Two-qubit gate error 0.4% | Circuit families it can run at 20 qubits 7 of 9 | Where a run is sold | |
| What it is 12 trapped ions; all-to-all | Two-qubit gate error 1.9% | Circuit families it can run at 20 qubits 0 of 9 | Where a run is sold | |
| What it is 98 trapped ions; all-to-all | Two-qubit gate error 0.079% | Circuit families it can run at 20 qubits 9 of 9 | Where a run is sold no public rate | |
| What it is 56 trapped ions; all-to-all | Two-qubit gate error 0.11% | Circuit families it can run at 20 qubits 9 of 9 | Where a run is sold no public rate | |
| What it is 56 trapped ions; all-to-all | Two-qubit gate error 0.083% | Circuit families it can run at 20 qubits 9 of 9 | Where a run is sold no public rate | |
| What it is 12 photonic qubits; 24 optical modes | Two-qubit gate error 1% | Circuit families it can run at 20 qubits 0 of 9 | Where a run is sold | |
| What it is 256 neutral atoms; analog only | Two-qubit gate error not published | Circuit families it can run at 20 qubits analog machine | Where a run is sold | |
| What it is 100 neutral atoms; analog only | Two-qubit gate error not published | Circuit families it can run at 20 qubits analog machine | Where a run is sold no public rate | |
| What it is Nighthawk r2; 120 qubits; square lattice | Two-qubit gate error 0.19% | Circuit families it can run at 20 qubits 7 of 9 | Where a run is sold | |
| What it is Heron r3; 156 qubits; heavy-hex | Two-qubit gate error 0.15% | Circuit families it can run at 20 qubits 8 of 9 | Where a run is sold | |
| What it is Heron r2; 156 qubits; heavy-hex | Two-qubit gate error 0.19% | Circuit families it can run at 20 qubits 7 of 9 | Where a run is sold | |
| What it is 108 qubits; square lattice | Two-qubit gate error 1.4% | Circuit families it can run at 20 qubits 0 of 9 | Where a run is sold | |
| What it is 54 qubits; square lattice | Two-qubit gate error 0.34% | Circuit families it can run at 20 qubits 5 of 9 | Where a run is sold | |
| What it is 20 qubits; square lattice | Two-qubit gate error 0.5% | Circuit families it can run at 20 qubits 1 of 9 | Where a run is sold |
A missing figure is one the vendor does not publish: – and "not published" are never a zero. Every published rate is on one page.
How this was measured
One host, one harness
Every simulator ran the same circuits on one host, single-threaded and core-pinned, driven through one common harness, so any overhead the harness adds is charged to the simulator. Each result was checked against a reference implementation: a simulator that ran fast but returned the wrong answer is left out of the ranking. Nothing is interpolated or rescaled. A time was measured, or it is not shown.
Run 2026-08-10, 8 cores, threads=1. numpy 2.4.6, scipy 1.18.0, torch 2.13.0, numba 0.66.0, pennylane 0.45.1, pennylane_lightning 0.45.0, qiskit 2.5.1, qulacs 0.6.14, tensorcircuit 1.8.0, cudaq 0.15.1. One host and one thread make the comparison reproducible. It is not a prediction for your laptop.
The interactive tool shows the same measurements for any workload you pick, with the quantum hardware that could run it and what a run would cost.