Gwil

Google Willow: Advancing Quantum Error Correction and Sampling with 105 Superconducting Qubits

Google Willow Research prototype

Google's Willow processor, with its 105 superconducting qubits, is a research prototype pushing the boundaries of quantum error correction and demonstrating unprecedented speed in random circuit sampling.

Google Quantum AI Superconducting transmon qubits Research prototype Physical qubits confidence: high

The Google Willow quantum processor, unveiled in 2024, represents a significant stride in the development of superconducting quantum computing. As a research prototype from Google Quantum AI, Willow is not merely an incremental increase in qubit count but a dedicated platform designed to tackle two of the most critical challenges in quantum computing: achieving robust quantum error correction (QEC) and demonstrating practical quantum advantage through random circuit sampling (RCS). With 105 physical superconducting transmon qubits, Willow builds upon Google's extensive experience in the field, notably following the Sycamore processor's quantum supremacy demonstration.

From a data analyst's perspective, Willow's importance lies in its concrete, measurable advancements. It provides crucial data points for understanding the current state-of-the-art in qubit fidelity, connectivity, and the efficacy of error mitigation and correction techniques. The system's architecture, a 2D grid with an average connectivity of 3.47 (typically 4-way), is optimized for the intricate qubit interactions required for both complex quantum algorithms and the overhead of error correction codes. Its status as an internal research-only system, developed and operated within Google's Santa Barbara, CA facility, underscores its role as a foundational testbed rather than a commercial offering.

Willow's development strategy is particularly noteworthy, featuring two distinct chip variants: one specifically optimized for quantum error correction research and another tailored for high-speed random circuit sampling. This dual-focus approach allows Google to explore different facets of quantum computing's potential simultaneously, gathering specialized data for each domain. For QEC, the goal is to demonstrate logical qubit performance that surpasses physical qubit error rates, a critical prerequisite for fault-tolerant quantum computing. For RCS, the objective is to further solidify the evidence for quantum computational advantage on specific, classically hard problems, pushing the boundaries of what's computationally feasible.

The metrics reported for Willow, such as single-qubit gate errors as low as 0.035% and two-qubit gate errors (CZ and iSWAP-like) at 0.33% and 0.14% respectively, are vital for assessing its performance against other superconducting platforms and alternative quantum technologies. These figures, alongside measurement fidelities and T1 coherence times, provide a comprehensive picture of the hardware's quality. Willow's achievements, including an XEB fidelity of 0.1% at 103 qubits for circuits of depth 40 and a reported 10^25x speedup over classical methods for RCS, are not just headline numbers; they are empirical data points that inform the entire quantum computing ecosystem about the current capabilities and future trajectory of the field. This profile aims to dissect these facts, offering a clear, metrics-driven overview for those analyzing the quantum hardware landscape.

Key metrics

Physical qubits
105 (2024)
Number of physical superconducting qubits
Benchmark headline
0.1
XEB fidelity 0.1% at 103 qubits depth 40 (2024) | RCS 10^25x faster than classical | Error correction Λ=2.14
Error-correction readiness
50/100
Heuristic score from topology + mode + error hints
Historical importance
0/100
Heuristic score from milestones + roadmap language
Native gates
SQ rotations | CZ | iSWAP-like | Multi-level reset
Gate alphabet you compile to
Connectivity
2D grid average connectivity 3.47 (4-way typical)
Mapping overhead + routing depth sensitivity

Technical specifications

Spec Details
System ID GWIL
Vendor Google Quantum AI
Technology Superconducting transmon qubits
Status Research prototype
Primary metric Physical qubits
Metric meaning Number of physical superconducting qubits
Qubit mode Transmon qubits for gate-based computation
Connectivity 2D grid average connectivity 3.47 (4-way typical)
Native gates SQ rotations | CZ | iSWAP-like | Multi-level reset
Error rates & fidelities SQ gate error 0.035% (2024) | TQ CZ 0.33% | TQ iSWAP 0.14% | Measurement 0.67-0.77% | T1 68-98us
Benchmarks XEB fidelity 0.1% at 103 qubits depth 40 (2024) | RCS 10^25x faster than classical | Error correction Λ=2.14
How to access Internal research only
Platforms Google Quantum AI internal
SDKs Cirq
Regions Santa Barbara CA
Account requirements Google collaborator
Pricing model Not applicable
Example prices Not applicable
Free tier / credits Not applicable
First announced 2024-12-09
First available 2024
Major revisions Chip 1 for QEC | Chip 2 for RCS
Retired / roadmap Active, roadmap to larger error-corrected systems
Notes Two chips with different optimizations

System profile

Qubit Technology and Architecture: Google Willow employs 105 physical superconducting transmon qubits, a technology choice that Google has consistently refined over several generations of processors. Transmon qubits are favored for their relatively long coherence times and ease of control, making them a leading candidate for large-scale quantum computing. The qubits are arranged in a 2D grid topology, which is a common and scalable architecture for superconducting circuits. This grid features an average connectivity of 3.47, with typical qubits having 4-way connectivity to their nearest neighbors. This level of connectivity is crucial for implementing complex quantum circuits and for the intricate entanglement operations required by many quantum error correction codes, allowing for efficient information transfer between adjacent qubits without excessive SWAP gates that can introduce additional errors.

Native Gate Set: The operational capabilities of Willow are defined by its native gate set, which includes single-qubit (SQ) rotations, controlled-Z (CZ) gates, iSWAP-like gates, and multi-level reset operations. Single-qubit rotations allow for arbitrary manipulation of individual qubit states. The CZ gate is a fundamental two-qubit entangling gate, essential for building complex quantum algorithms. The iSWAP-like gate offers an alternative entangling operation, often providing different error characteristics or being more efficient for specific types of interactions. The multi-level reset capability is important for quickly preparing qubits in a known initial state, which is critical for iterative algorithms and error correction cycles, where qubits need to be reset after measurement or error detection.

Performance Metrics: Error Rates and Coherence: The fidelity of quantum operations is paramount, and Willow reports impressive figures for a system of its scale. Single-qubit gate errors are measured at a remarkably low 0.035% (0.99965 fidelity) in 2024. Two-qubit gate errors, which are typically more challenging to achieve high fidelity for, are reported at 0.33% for CZ gates and 0.14% for iSWAP-like gates. These figures are competitive within the superconducting qubit landscape and indicate a high level of control over qubit interactions. Measurement error rates range from 0.67% to 0.77%, which is a critical factor for the accuracy of readout and the effectiveness of error correction. Coherence times, specifically the T1 relaxation time, are reported between 68-98 microseconds. While these times are relatively short compared to some other quantum modalities like trapped ions, they are sufficient for executing circuits of significant depth given the speed of superconducting gates, allowing for many gate operations within the coherence window.

Benchmarking Achievements: Willow has demonstrated several key benchmarks that highlight its advanced capabilities. It achieved an XEB (Cross-Entropy Benchmarking) fidelity of 0.1% on 103 qubits for circuits of depth 40 in 2024. XEB is a robust method for validating the performance of quantum processors on complex, random circuits, and this result indicates a high degree of control and low error accumulation across a large number of qubits and gate operations. Furthermore, Willow has been used to demonstrate random circuit sampling (RCS) with a reported speedup of 10^25x compared to classical supercomputers, reinforcing the concept of quantum computational advantage for specific, highly complex tasks. Crucially for the future of quantum computing, Willow has also shown significant progress in quantum error correction, achieving a logical error suppression metric (Λ) of 2.14. This indicates that the system is capable of reducing the effective error rate of a logical qubit below that of its constituent physical qubits, a foundational step towards fault-tolerant quantum computing. The ability to achieve error correction 'below threshold' is a major milestone, suggesting that increasing the number of physical qubits could lead to further reductions in logical error rates.

System Limitations and Practical Considerations: As a research prototype, Willow's current operational limits are primarily defined by its experimental nature. While it can execute circuits up to depth 40, specific limits on the number of shots per circuit, queueing policies, or total daily runtime are not publicly confirmed, reflecting its internal research focus. The system's design inherently involves tradeoffs, such as balancing qubit coherence with the need for robust shielding and control wiring. While its error rates are impressive for a NISQ (Noisy Intermediate-Scale Quantum) device, they are still considered high for achieving full fault-tolerant quantum computing (FTQC) without substantial error correction overhead. However, its speed and gate fidelities position it favorably against slower, albeit potentially more coherent, technologies like trapped ions for certain types of computations. The system is exclusively accessible internally within Google Quantum AI, primarily through the Cirq SDK, and requires Google collaborator status for access, reinforcing its role as a dedicated research instrument rather than a general-purpose cloud offering.

Generation lineage (family-level)
Heuristic chain based on common naming. Verify by revision/date for strict claims.
Related systems (same vendor)
Cross-system comparison (same vendor)
System Status Primary metric
Google Sycamore Research prototype Physical qubits: 53 (2019)
Google Surface-code logical prototype Research prototype Logical qubits: 1 (2023)

Access & pricing

How you access it
  • <b>Public Access:</b> Currently, Google Willow is not publicly accessible.
  • <b>How to Access:</b> Access is restricted to internal research within Google Quantum AI.
  • <b>Platforms:</b> The system operates exclusively on Google Quantum AI's internal platforms.
  • <b>Geographic Availability:</b> The hardware is located and operated in Santa Barbara, CA.
  • <b>Software Development Kits (SDKs):</b> Research is conducted using Google's Cirq SDK.
  • <b>Account Requirements:</b> Access is limited to Google collaborators and internal research teams.
  • <b>Access Notes:</b> As a research prototype, Willow's primary purpose is internal experimentation and development, not external user access.
How costs sneak up
  • <b>Public Pricing:</b> There is no public pricing model for Google Willow.
  • <b>Pricing Model:</b> Not applicable, as the system is for internal research only.
  • <b>Example Prices:</b> No example prices are available.
  • <b>Cost Drivers:</b> Not applicable for external users; internal costs are part of Google's R&D budget.
  • <b>Free Tier/Credits:</b> No free tier or credits are offered, as it's not a commercial service.
  • <b>Pricing Notes:</b> Willow is exclusively a research and development platform, therefore commercial pricing models do not apply.

Status timeline

The journey of Google Willow began with its official announcement on December 9, 2024, marking a significant milestone in Google Quantum AI's roadmap towards fault-tolerant quantum computing. While the announcement solidified its public presence, the underlying development and initial availability of the system for internal research commenced earlier in 2024. Willow represents a direct evolution from Google's prior quantum processors, most notably building upon the foundational work established by the Sycamore processor, which famously demonstrated quantum supremacy in 2019. Sycamore proved the potential of superconducting qubits for complex computations, and Willow aims to translate that potential into practical, error-corrected quantum computation.

A key strategic decision in Willow's development, and a major revision from prior single-purpose chips, was the creation of two distinct chip variants. One variant, designated 'Chip 1,' was specifically engineered and optimized for quantum error correction (QEC) research. This specialization allowed Google's researchers to meticulously design and test error correction codes, focusing on the intricate qubit interactions and measurement feedback loops necessary to protect quantum information from noise. The other variant, 'Chip 2,' was optimized for random circuit sampling (RCS), pushing the boundaries of computational advantage demonstrations. This dual-chip approach enabled parallel advancements in both fundamental error mitigation and the practical demonstration of quantum computational power, providing targeted data for each research area.

Throughout 2024, Willow achieved critical benchmarks that underscore its importance. These included demonstrating an XEB fidelity of 0.1% on 103 qubits at a circuit depth of 40, a testament to the system's overall performance and control. More profoundly, Willow showcased significant progress in quantum error correction, achieving logical error suppression with a Λ (Lambda) value of 2.14. This result is crucial because it indicates that the logical qubit's error rate is being reduced relative to its physical constituents, a necessary condition for scaling towards fault-tolerant quantum computers. The RCS variant also delivered a staggering 10^25x speedup over classical methods, further solidifying the case for quantum advantage in specific computational tasks.

Looking ahead, the roadmap for Google Willow is firmly active and focused on the development of larger, more robust error-corrected systems. Willow is not intended as a final product but as a critical stepping stone. Its research findings, particularly in error correction, are directly informing the design and engineering of future generations of Google's quantum processors. The insights gained from Willow's dual-chip approach and its performance metrics are invaluable for understanding the tradeoffs inherent in quantum hardware design and for charting a clear path towards truly fault-tolerant quantum computing. The continuous development within Google's Santa Barbara fab ensures a rapid iteration cycle, with Willow's legacy being the foundational data and techniques that will enable the next era of quantum computation.

What to verify next

  • <b>Production Scaling:</b> Monitor Google's announcements for evidence of scaling Willow's architecture to even larger qubit counts, particularly how error rates and coherence scale with system size.
  • <b>Updated Benchmarks:</b> Look for new benchmark results, especially those demonstrating improved XEB fidelities, deeper circuits, or enhanced error correction performance (e.g., higher Λ values or demonstration of more complex error correction codes).
  • <b>Logical Qubit Performance:</b> Verify if Google provides further details on logical qubit performance, such as the lifetime of encoded quantum information or the fidelity of logical gates, beyond the initial error suppression metrics.
  • <b>Application-Specific Benchmarks:</b> Investigate if Willow is applied to more domain-specific benchmarks beyond RCS, such as simulations of molecular systems or optimization problems, to assess its utility for practical applications.
  • <b>Coherence Improvements:</b> Track any reported improvements in qubit coherence times (T1, T2) or reductions in gate operation times, as these directly impact the achievable circuit depth and overall computational power.
  • <b>Public Access Plans:</b> While currently internal, any future indications or announcements regarding potential public access or cloud integration for Willow or its successor systems would be a critical development to verify.

FAQ

What is Google Willow?

Google Willow is a 105-qubit superconducting quantum processor developed by Google Quantum AI. Announced in 2024, it serves as a research prototype focused on advancing quantum error correction and demonstrating quantum computational advantage through random circuit sampling.

How many qubits does Google Willow have and what type are they?

Google Willow features 105 physical qubits. These are superconducting transmon qubits, a leading technology in the field known for its balance of coherence and controllability.

What are Willow's primary research objectives?

Willow has two primary objectives: first, to conduct advanced research into quantum error correction (QEC), aiming to demonstrate logical qubit performance superior to physical qubits. Second, to further explore quantum computational advantage through high-speed random circuit sampling (RCS), building on previous 'quantum supremacy' demonstrations.

Is Google Willow publicly accessible?

No, Google Willow is not publicly accessible. It is an internal research prototype used exclusively by Google Quantum AI collaborators and researchers at their Santa Barbara, CA facility.

What are some key performance metrics for Willow?

Key performance metrics include a single-qubit gate error rate of 0.035%, two-qubit CZ gate error of 0.33%, and iSWAP-like gate error of 0.14%. It achieved an XEB fidelity of 0.1% on 103 qubits at depth 40 and demonstrated a 10^25x speedup for random circuit sampling over classical methods. T1 coherence times range from 68-98 microseconds.

What is the significance of Willow's error correction benchmarks?

Willow achieved a logical error suppression metric (Λ) of 2.14, which is highly significant. This indicates that the system can reduce the effective error rate of a logical qubit below that of its constituent physical qubits, a crucial step towards building fault-tolerant quantum computers that can perform complex calculations reliably.

What software development kit (SDK) is used with Google Willow?

Internal research on Google Willow primarily utilizes Google's Cirq SDK, a Python library designed for writing, manipulating, and optimizing quantum circuits.

What is the future roadmap for Google Willow?

Willow is an active research platform, and its roadmap is focused on paving the way for larger, more robust error-corrected quantum systems. The insights gained from Willow's experiments, particularly in error correction and scaling, will directly inform the design and development of Google's next-generation quantum processors.



Also in Quantum Computing

Xanadu X8

Xanadu X12

Xanadu Borealis

Subscribe