All results measured on IBM ibm_fez (Heron processor)

Real Hardware. Real Results.

Every number below was measured on IBM quantum hardware, not simulators. We publish everything, including where we don't improve.

2.7×

Peak Fidelity Gain

94%

Max Error Reduction

$199

Per Month, Starting

~30 s

Per Correction

DISTRIBUTION CORRECTION

Output Distribution Fidelity

GHZ entanglement circuits corrected on IBM ibm_fez. Fidelity improvement measures how much closer the corrected output distribution is to the ideal quantum state.

GHZ-20

20 qubits · NEC

+171%

2.71×

20-qubit GHZ entanglement circuit

GHZ-50

50 qubits · NEC

+130%

2.30×

50-qubit GHZ entanglement circuit

GHZ-30

30 qubits · NEC

+98%

1.98×

30-qubit GHZ entanglement circuit

What this means: A +171% fidelity improvement means the corrected result is 2.71× closer to the mathematically perfect output than the raw hardware result. These are measured on real IBM quantum processors, not simulators.

EXPECTATION VALUE CORRECTION

Production Circuit Results

VQE and QAOA circuits, the workloads quantum researchers actually run. Error reduction measures what fraction of the expectation value error is removed.

VQE: H₂ molecule

94%

expectation error removed

Qubits4

Circuit depth12

MethodML-CDR

Hydrogen molecule ground state energy estimation

QAOA: MaxCut

80%

expectation error removed

Qubits6

Circuit depth24

MethodO-CDR

6-node combinatorial graph optimization

VQE: 20-qubit

31%

expectation error removed

Qubits20

Circuit depth103

MethodO-CDR

Deep production-scale variational circuit

Why the range matters: Small circuits (4–6 qubits) show 80–94% error reduction. At 20 qubits and 103 layers of depth (a realistic production workload), we still reduce error by 31%. This scaling behavior is consistent across all error mitigation approaches.

90% of the Benefit. 1% of the Cost.

How Qubience compares to enterprise quantum error mitigation solutions, honestly.

Industry benchmark data from arxiv:2608.05202 (August 2026 independent benchmark). Enterprise pricing estimated from industry sources.

Solution	Error Reduction	Annual Cost	QPU Overhead	Setup Time
Leading Enterprise Vendor	4.7×	Custom (enterprise)	12–17×	Weeks
Enterprise Middleware Platform	3.1×	Custom (enterprise)	~1.5×	Days
QubienceYou are here	1.7–2.7×	$2,388	~1×	Minutes
Native Hardware Mitigation	1.1×	Free	~1.5–2×	Built-in
Open-Source Toolkit	1.5–2×	Free (DIY)	3–5×	Hours–Days

Transparent pricing from day one

Enterprise vendors require custom quotes and long contracts. Qubience starts at $199/mo with no surprises. Same quantum hardware, radically different model.

Zero additional QPU time

Enterprise solutions add 12–17× QPU overhead. Open-source toolkits add 3–5×. Qubience corrects your results without running a single extra circuit.

Minutes to integrate, not weeks

pip install qubience → 4 lines of code → corrected results. No consulting engagement required.

What Error Correction Actually Costs

For a research lab running 500 circuits per month, here's what one year of quantum error correction looks like.

Enterprise vendors

Custom enterprise contracts

Custom

per year

Per circuitContact vendor

Improvement3–5×

Enterprise contracts, custom integration, months of onboarding. Pricing not publicly available.

Best value

Qubience

Professional plan

$2,388

per year

Per circuit$0.40

Improvement1.7–2.7×

Self-serve, instant setup, corrected results in ~30 seconds. No hidden fees.

DIY Approach

Open-source toolkits

$0 + eng. time

per year

Per circuit$0 + 3–5× QPU

Improvement1.5–2×

Free software but significant engineering effort and 3–5× extra QPU time ($46–77 per job at IBM rates).

IBM QPU time costs $1.60/second on pay-as-you-go. A 12–17× overhead adds $300–470 per job. Qubience adds $0 in extra QPU time.

THE HIDDEN COST MOST VENDORS DON'T MENTION

QPU Time Is Expensive

IBM charges $1.60 per second of QPU time on pay-as-you-go. Methods that require extra circuit executions multiply that cost quickly.

Raw execution

$29

per job (~18s)

$0extra QPU cost

Qubience

$29

per job (~18s)

$0extra QPU cost

Open-Source ZNE (3×)

$86

per job (~54s)

+$57extra QPU cost

Enterprise (12–17×)

$400

per job (~250s)

+$371extra QPU cost

Based on IBM Quantum pay-as-you-go rate ($96/min). Flex plan: $72/min. Premium: $48/min. QPU time calculated from benchmarks in arxiv:2608.05202.

TRANSPARENCY

Our Approach to Benchmarking

What we tested

All circuits run on IBM ibm_fez (156-qubit Heron processor)
8,192 shots per circuit for statistical significance
Multiple correction methods tested (NEC, ML-CDR, O-CDR)
Best-performing method reported for each circuit type
Results reproducible with our public SDK

What we don't hide

Shallow circuits (<10 layers) show minimal improvement
Random circuits without systematic error patterns don't benefit
Error reduction scales inversely with qubit count on deep VQE
We are not as powerful as leading enterprise vendors on raw error reduction
Noisier hardware sees larger improvements than newer processors

Independent Benchmark Reference

Competitor performance data referenced from the independent quantum error mitigation benchmark paper published August 2026. We encourage you to read it and draw your own conclusions.

arxiv:2608.05202 — Benchmark of Quantum Error Mitigation

Start Correcting Quantum Errors Today

10 free API calls. No credit card required. Results in ~30 seconds.

pip install qubience · 4 lines of code · Corrected results

All results measured on IBM ibm_fez (Heron processor)

Real Hardware. Real Results.

Every number below was measured on IBM quantum hardware, not simulators. We publish everything, including where we don't improve.

2.7×

Peak Fidelity Gain

94%

Max Error Reduction

$199

Per Month, Starting

~30 s

Per Correction

DISTRIBUTION CORRECTION

Output Distribution Fidelity

GHZ entanglement circuits corrected on IBM ibm_fez. Fidelity improvement measures how much closer the corrected output distribution is to the ideal quantum state.

GHZ-20

20 qubits · NEC

+171%

2.71×

20-qubit GHZ entanglement circuit

GHZ-50

50 qubits · NEC

+130%

2.30×

50-qubit GHZ entanglement circuit

GHZ-30

30 qubits · NEC

+98%

1.98×

30-qubit GHZ entanglement circuit

EXPECTATION VALUE CORRECTION

Production Circuit Results

VQE and QAOA circuits, the workloads quantum researchers actually run. Error reduction measures what fraction of the expectation value error is removed.

VQE: H₂ molecule

94%

expectation error removed

Qubits4

Circuit depth12

MethodML-CDR

Hydrogen molecule ground state energy estimation

QAOA: MaxCut

80%

expectation error removed

Qubits6

Circuit depth24

MethodO-CDR

6-node combinatorial graph optimization

VQE: 20-qubit

31%

expectation error removed

Qubits20

Circuit depth103

MethodO-CDR

Deep production-scale variational circuit

90% of the Benefit. 1% of the Cost.

How Qubience compares to enterprise quantum error mitigation solutions, honestly.

Industry benchmark data from arxiv:2608.05202 (August 2026 independent benchmark). Enterprise pricing estimated from industry sources.

Solution	Error Reduction	Annual Cost	QPU Overhead	Setup Time
Leading Enterprise Vendor	4.7×	Custom (enterprise)	12–17×	Weeks
Enterprise Middleware Platform	3.1×	Custom (enterprise)	~1.5×	Days
QubienceYou are here	1.7–2.7×	$2,388	~1×	Minutes
Native Hardware Mitigation	1.1×	Free	~1.5–2×	Built-in
Open-Source Toolkit	1.5–2×	Free (DIY)	3–5×	Hours–Days

Transparent pricing from day one

Enterprise vendors require custom quotes and long contracts. Qubience starts at $199/mo with no surprises. Same quantum hardware, radically different model.

Zero additional QPU time

Enterprise solutions add 12–17× QPU overhead. Open-source toolkits add 3–5×. Qubience corrects your results without running a single extra circuit.

Minutes to integrate, not weeks

pip install qubience → 4 lines of code → corrected results. No consulting engagement required.

What Error Correction Actually Costs

For a research lab running 500 circuits per month, here's what one year of quantum error correction looks like.

Enterprise vendors

Custom enterprise contracts

Custom

per year

Per circuitContact vendor

Improvement3–5×

Enterprise contracts, custom integration, months of onboarding. Pricing not publicly available.

Best value

Qubience

Professional plan

$2,388

per year

Per circuit$0.40

Improvement1.7–2.7×

Self-serve, instant setup, corrected results in ~30 seconds. No hidden fees.

DIY Approach

Open-source toolkits

$0 + eng. time

per year

Per circuit$0 + 3–5× QPU

Improvement1.5–2×

Free software but significant engineering effort and 3–5× extra QPU time ($46–77 per job at IBM rates).

IBM QPU time costs $1.60/second on pay-as-you-go. A 12–17× overhead adds $300–470 per job. Qubience adds $0 in extra QPU time.

THE HIDDEN COST MOST VENDORS DON'T MENTION

QPU Time Is Expensive

IBM charges $1.60 per second of QPU time on pay-as-you-go. Methods that require extra circuit executions multiply that cost quickly.

Raw execution

$29

per job (~18s)

$0extra QPU cost

Qubience

$29

per job (~18s)

$0extra QPU cost

Open-Source ZNE (3×)

$86

per job (~54s)

+$57extra QPU cost

Enterprise (12–17×)

$400

per job (~250s)

+$371extra QPU cost

Based on IBM Quantum pay-as-you-go rate ($96/min). Flex plan: $72/min. Premium: $48/min. QPU time calculated from benchmarks in arxiv:2608.05202.

TRANSPARENCY

Our Approach to Benchmarking

What we tested

All circuits run on IBM ibm_fez (156-qubit Heron processor)
8,192 shots per circuit for statistical significance
Multiple correction methods tested (NEC, ML-CDR, O-CDR)
Best-performing method reported for each circuit type
Results reproducible with our public SDK

What we don't hide

Shallow circuits (<10 layers) show minimal improvement
Random circuits without systematic error patterns don't benefit
Error reduction scales inversely with qubit count on deep VQE
We are not as powerful as leading enterprise vendors on raw error reduction
Noisier hardware sees larger improvements than newer processors

Independent Benchmark Reference

Competitor performance data referenced from the independent quantum error mitigation benchmark paper published August 2026. We encourage you to read it and draw your own conclusions.

arxiv:2608.05202 — Benchmark of Quantum Error Mitigation

Start Correcting Quantum Errors Today

10 free API calls. No credit card required. Results in ~30 seconds.

pip install qubience · 4 lines of code · Corrected results