MaxCut on a ring: which simulator is fastest

QAOA over a cycle graph

8 simulators ran this circuit at 4 to 24 qubits, 27 workloads in all. Qulacs was the fastest in 14, Google qsim in 9, Qiskit Aer (MPS) in 2, TensorCircuit in 1 and PennyLane Lightning in 1. The tables below hold every time as it was measured, with the fastest in each row marked. Open this circuit in the interactive tool to see the hardware that could run it and what a run would cost.

Expectation value, local observable, depth 3

Time per run, by qubit count

QubitsPennyLanePennyLane LightningQiskitQiskit AerQiskit Aer (MPS)Google qsimQulacsTensorCircuit
4 qubits
PennyLane
5.17 ms
PennyLane Lightning
2.84 ms
Qiskit
2.01 ms
Qiskit Aer
1.56 ms
Qiskit Aer (MPS)
1.84 ms
Google qsim
2.67 ms
Qulacs
63.1 µs
TensorCircuit
199 µs
8 qubits
PennyLane
10.6 ms
PennyLane Lightning
4.94 ms
Qiskit
4.71 ms
Qiskit Aer
2.71 ms
Qiskit Aer (MPS)
4.28 ms
Google qsim
5.15 ms
Qulacs
154 µs
TensorCircuit
387 µs
12 qubits
PennyLane
17.8 ms
PennyLane Lightning
6.97 ms
Qiskit
9.95 ms
Qiskit Aer
5.99 ms
Qiskit Aer (MPS)
19.9 ms
Google qsim
7.52 ms
Qulacs
1.14 ms
TensorCircuit
1.43 ms
16 qubits
PennyLane
82.9 ms
PennyLane Lightning
14.2 ms
Qiskit
105 ms
Qiskit Aer
30.4 ms
Qiskit Aer (MPS)
54.6 ms
Google qsim
11.0 ms
Qulacs
20.9 ms
TensorCircuit
21.2 ms
20 qubits
PennyLane
1.82 s
PennyLane Lightning
176 ms
Qiskit
2.86 s
Qiskit Aer
517 ms
Qiskit Aer (MPS)
88.7 ms
Google qsim
41.3 ms
Qulacs
433 ms
TensorCircuit
596 ms
24 qubits
PennyLane
skip
PennyLane Lightning
6.43 s
Qiskit
skip
Qiskit Aer
9.94 s
Qiskit Aer (MPS)
123 ms
Google qsim
1.07 s
Qulacs
10.6 s
TensorCircuit
12.8 s

Expectation value, global observable (all_z), depth 3

Time per run, by qubit count

QubitsPennyLanePennyLane LightningQiskitQiskit AerQiskit Aer (MPS)Google qsimQulacsTensorCircuit
4 qubits
PennyLane
5.56 ms
PennyLane Lightning
3.02 ms
Qiskit
2.13 ms
Qiskit Aer
1.62 ms
Qiskit Aer (MPS)
1.95 ms
Google qsim
2.74 ms
Qulacs
72.0 µs
TensorCircuit
207 µs
8 qubits
PennyLane
11.3 ms
PennyLane Lightning
5.23 ms
Qiskit
4.72 ms
Qiskit Aer
2.99 ms
Qiskit Aer (MPS)
4.93 ms
Google qsim
5.23 ms
Qulacs
197 µs
TensorCircuit
428 µs
12 qubits
PennyLane
19.9 ms
PennyLane Lightning
7.65 ms
Qiskit
10.5 ms
Qiskit Aer
6.66 ms
Qiskit Aer (MPS)
32.2 ms
Google qsim
7.74 ms
Qulacs
1.91 ms
TensorCircuit
2.01 ms
16 qubits
PennyLane
95.7 ms
PennyLane Lightning
18.2 ms
Qiskit
107 ms
Qiskit Aer
38.1 ms
Qiskit Aer (MPS)
131 ms
Google qsim
12.0 ms
Qulacs
36.9 ms
TensorCircuit
58.3 ms
20 qubits
PennyLane
2.05 s
PennyLane Lightning
285 ms
Qiskit
2.81 s
Qiskit Aer
672 ms
Qiskit Aer (MPS)
262 ms
Google qsim
55.1 ms
Qulacs
740 ms
TensorCircuit
1.66 s
24 qubits
PennyLane
skip
PennyLane Lightning
12.2 s
Qiskit
skip
Qiskit Aer
12.8 s
Qiskit Aer (MPS)
424 ms
Google qsim
1.48 s
Qulacs
16.6 s
TensorCircuit
42.9 s

Expectation value, global observable (pauli_mix), depth 3

Time per run, by qubit count

QubitsPennyLanePennyLane LightningQiskitQiskit AerQiskit Aer (MPS)Google qsimQulacsTensorCircuit
8 qubits
PennyLane
12.7 ms
PennyLane Lightning
5.59 ms
Qiskit
5.70 ms
Qiskit Aer
3.01 ms
Qiskit Aer (MPS)
5.72 ms
Google qsim
5.68 ms
Qulacs
234 µs
TensorCircuit
541 µs
12 qubits
PennyLane
22.7 ms
PennyLane Lightning
8.47 ms
Qiskit
12.4 ms
Qiskit Aer
7.21 ms
Qiskit Aer (MPS)
56.4 ms
Google qsim
8.49 ms
Qulacs
2.61 ms
TensorCircuit
2.90 ms
16 qubits
PennyLane
116 ms
PennyLane Lightning
27.1 ms
Qiskit
121 ms
Qiskit Aer
47.7 ms
Qiskit Aer (MPS)
274 ms
Google qsim
15.9 ms
Qulacs
52.1 ms
TensorCircuit
96.2 ms
20 qubits
PennyLane
2.39 s
PennyLane Lightning
543 ms
Qiskit
2.83 s
Qiskit Aer
858 ms
Qiskit Aer (MPS)
skip
Google qsim
151 ms
Qulacs
1.03 s
TensorCircuit
2.50 s

Expectation value, global observable (zz_sum), depth 3

Time per run, by qubit count

QubitsPennyLanePennyLane LightningQiskitQiskit AerQiskit Aer (MPS)Google qsimQulacsTensorCircuit
8 qubits
PennyLane
11.7 ms
PennyLane Lightning
5.43 ms
Qiskit
5.27 ms
Qiskit Aer
2.92 ms
Qiskit Aer (MPS)
5.03 ms
Google qsim
5.52 ms
Qulacs
191 µs
TensorCircuit
502 µs
12 qubits
PennyLane
20.2 ms
PennyLane Lightning
7.92 ms
Qiskit
11.8 ms
Qiskit Aer
6.85 ms
Qiskit Aer (MPS)
35.0 ms
Google qsim
8.38 ms
Qulacs
1.95 ms
TensorCircuit
2.53 ms
16 qubits
PennyLane
93.4 ms
PennyLane Lightning
20.6 ms
Qiskit
113 ms
Qiskit Aer
39.4 ms
Qiskit Aer (MPS)
141 ms
Google qsim
13.2 ms
Qulacs
38.6 ms
TensorCircuit
77.6 ms
20 qubits
PennyLane
1.88 s
PennyLane Lightning
332 ms
Qiskit
2.84 s
Qiskit Aer
685 ms
Qiskit Aer (MPS)
skip
Google qsim
87.2 ms
Qulacs
762 ms
TensorCircuit
2.25 s

Gradient, local observable, depth 3

Time per run, by qubit count

QubitsPennyLanePennyLane LightningQiskitQiskit AerQiskit Aer (MPS)Google qsimQulacsTensorCircuit
4 qubits
PennyLane
23.6 ms
PennyLane Lightning
5.73 ms
Qiskit
unsupported
Qiskit Aer
unsupported
Qiskit Aer (MPS)
unsupported
Google qsim
unsupported
Qulacs
255 µs
TensorCircuit
372 µs
8 qubits
PennyLane
48.6 ms
PennyLane Lightning
9.69 ms
Qiskit
unsupported
Qiskit Aer
unsupported
Qiskit Aer (MPS)
unsupported
Google qsim
unsupported
Qulacs
681 µs
TensorCircuit
858 µs
12 qubits
PennyLane
113 ms
PennyLane Lightning
14.9 ms
Qiskit
unsupported
Qiskit Aer
unsupported
Qiskit Aer (MPS)
unsupported
Google qsim
unsupported
Qulacs
6.05 ms
TensorCircuit
4.31 ms
16 qubits
PennyLane
764 ms
PennyLane Lightning
61.5 ms
Qiskit
unsupported
Qiskit Aer
unsupported
Qiskit Aer (MPS)
unsupported
Google qsim
unsupported
Qulacs
120 ms
TensorCircuit
104 ms

Statevector, global observable (all_z), depth 3

Time per run, by qubit count

QubitsPennyLanePennyLane LightningQiskitQiskit AerQiskit Aer (MPS)Google qsimQulacsTensorCircuit
8 qubits
PennyLane
10.9 ms
PennyLane Lightning
7.03 ms
Qiskit
4.86 ms
Qiskit Aer
2.93 ms
Qiskit Aer (MPS)
4.80 ms
Google qsim
5.15 ms
Qulacs
155 µs
TensorCircuit
unsupported
12 qubits
PennyLane
19.1 ms
PennyLane Lightning
9.39 ms
Qiskit
11.1 ms
Qiskit Aer
6.43 ms
Qiskit Aer (MPS)
23.8 ms
Google qsim
7.61 ms
Qulacs
1.13 ms
TensorCircuit
unsupported
16 qubits
PennyLane
88.1 ms
PennyLane Lightning
17.3 ms
Qiskit
113 ms
Qiskit Aer
31.0 ms
Qiskit Aer (MPS)
123 ms
Google qsim
12.0 ms
Qulacs
21.0 ms
TensorCircuit
unsupported

Which quantum computers can run it

At depth 3, 1,000 shots

Quantum computerCan run it up toBorderline up toEstimated cost at that size
Can run it up to
8 qubits
Borderline up to
32 qubits
Estimated cost at that size
$200.30–$600.90
via AWS Braket
Can run it up to
20 qubits
Borderline up to
32 qubits
Estimated cost at that size
$200.30–$600.90
via AWS Braket
Can run it up to
4 qubits
Borderline up to
12 qubits
Estimated cost at that size
$23.80–$71.40
via AWS Braket
Can run it up to
96 qubits
Borderline up to
–
Estimated cost at that size
Not priced
this vendor publishes no rate card
Can run it up to
48 qubits
Borderline up to
–
Estimated cost at that size
Not priced
this vendor publishes no rate card
Can run it up to
48 qubits
Borderline up to
–
Estimated cost at that size
Not priced
this vendor publishes no rate card
Can run it up to
–
Borderline up to
–
Estimated cost at that size
–
Can run it up to
–
Borderline up to
–
Estimated cost at that size
–
Can run it up to
–
Borderline up to
–
Estimated cost at that size
–
Can run it up to
24 qubits
Borderline up to
96 qubits
Estimated cost at that size
from $0.015
via IBM Quantum. 0.01s at maximum throughput, so this is a floor
Can run it up to
32 qubits
Borderline up to
128 qubits
Estimated cost at that size
from $0.421
via IBM Quantum. 0.26s at maximum throughput, so this is a floor
Can run it up to
24 qubits
Borderline up to
96 qubits
Estimated cost at that size
from $0.421
via IBM Quantum. 0.26s at maximum throughput, so this is a floor
Can run it up to
–
Borderline up to
24 qubits
Estimated cost at that size
$0.725–$2.17
via AWS Braket
Can run it up to
16 qubits
Borderline up to
48 qubits
Estimated cost at that size
$1.90–$5.70
via AWS Braket
Can run it up to
8 qubits
Borderline up to
20 qubits
Estimated cost at that size
$1.75–$5.25
via AWS Braket

Sizes are the grid's own: 4, 8, 12, 16, 20, 24, 32, 48, 64, 96, 128, 156 qubits. A device can run a circuit when a clean run is more likely than not, and is borderline when it is better than one in fifty. The cost is for one run of 1,000 shots at the largest size in the row, as a band and not a quote.

The simulators

Each one across every circuit family

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.

Hardware figures each carry the source they came from and the date they were measured. The oldest in this set is 626 days old. Costs are bands, not quotes.

Host: Linux-6.8.0-1051-aws-x86_64-with-glibc2.35, 8 cores, threads=1, core-pinned. 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. Every simulator was driven through one common harness, so any harness binding overhead is charged to the simulator rather than hidden.

© 2026 EigenOrbit. All rights reserved.