Pasqal has achieved an industry first in neutral atom quantum computing, demonstrating that its logical qubits outperform standard physical qubits when solving complex differential equations on actual hardware. These equations, crucial for modeling everything from energy systems to financial markets, are notoriously difficult to solve accurately and serve as a key benchmark for advanced computing. In a recent study published on arXiv, Pasqal’s quantum processor improved solution accuracy by more than 50% on average, and achieved a ten-fold increase in speed on certain challenging problems, despite the increased complexity of the logical-qubit method. “What this work demonstrates is that logical qubits are not only theoretically preferable—they are already performing better on a real computational task,” said Loïc Henriet, Chief Technology Officer at Pasqal. This advancement offers concrete evidence that logical qubits can overcome error limitations hindering the progress of quantum computing.
Logical Qubits Outperform Physical Qubits Solving Differential Equations
Pasqal has demonstrated a significant improvement in speed and accuracy, showcasing the power of logical qubits over their physical counterparts when tackling complex calculations. The French quantum computing firm recently revealed research, published on arXiv, detailing how its logical qubit approach significantly outperformed conventional methods in solving differential equations, a benchmark for advanced computing systems. This achievement isn’t merely theoretical; Pasqal deployed its quantum processor to deliver these results on actual hardware, a crucial step toward practical quantum utility. The study focused on differential equations, mathematical formulations essential for modeling diverse phenomena from energy grids to financial markets. On particularly challenging nonlinear problems, the improvement reached a factor of ten. This performance boost is noteworthy given the increased complexity of the logical qubit method itself, highlighting its practical value despite the added computational overhead.
Pasqal’s success builds upon a foundation of analog quantum computing, but represents a deliberate shift toward fault tolerance. The company’s neutral-atom processor currently achieves a gate fidelity of 99.4%, a level of performance that enabled these application-grade results. Researchers implemented a quantum kernel algorithm, comparing the performance of both physical and logical qubits across a dataset of 1,000 differential equations.
The logical implementation utilized a more intricate circuit, encoding two logical qubits into four physical qubits using a quantum error-detecting code, yet still yielded demonstrably more accurate solutions. Henriet added that researchers were surprised to find their logical qubits were naturally resistant to the types of noise that make solving differential equations harder. They achieved better results than initially anticipated, and running complete applications reveals insights that testing individual building blocks alone cannot. This research is a direct outcome of the PROQCIMA programme, a French initiative fostering advancements in fault-tolerant quantum computing and logical qubit architectures.
99.4% Gate Fidelity Achieved on Neutral-Atom Processor
Pasqal’s recent advancements are reshaping quantum computation, moving beyond theoretical demonstrations toward practical applications with demonstrably improved performance. While many quantum computing approaches struggle with maintaining qubit stability, Pasqal has focused on building logical qubits, encoded units designed to actively mitigate errors, and has now demonstrated their superiority over standard physical qubits in solving complex problems. This isn’t simply a matter of theoretical advantage; the company has achieved a 99.4% combined gate fidelity on its neutral-atom processor, a crucial step toward reliable quantum calculations. In a study published on arXiv, researchers utilized Pasqal’s quantum processor to solve these notoriously difficult equations, achieving more than 50% improvement in accuracy on average. The team systematically compared results across 1,000 equations, implementing a quantum kernel algorithm at both the physical and logical qubit levels.
This achievement extends beyond isolated subroutines; Pasqal ran a complete application, allowing them to pinpoint the most impactful error sources for this specific class of problem. This granular understanding is now informing the next generation of hardware development. The 99.4% combined gate fidelity of Pasqal’s neutral-atom processor was a prerequisite for achieving these application-grade results, showcasing a deliberate progression from analog quantum computing toward fault tolerance. The research, funded by the PROQCIMA program, highlights the benefits of a collaborative approach between French academia and industry. The ability to solve these complex equations with greater accuracy, operating as accelerators within hybrid quantum-classical workflows, represents a significant stride toward realizing the potential of quantum utility at scale.
What this work demonstrates is that logical qubits are not only theoretically preferable – they are already performing better on a real computational task,” said Loïc Henriet, Chief Technology Officer at Pasqal.
Loïc Henriet, Chief Technology Officer at Pasqal
Quantum Kernel Algorithm Benchmarks Across 1,000 Equations
Pasqal, a leader in neutral atom quantum computing, is demonstrating tangible progress in error mitigation with a recent benchmark involving the solution of 1,000 differential equations. The company’s approach, utilizing logical qubits, has achieved a significant performance leap over conventional methods and standard physical qubits, marking an industry first for this quantum computing architecture. This isn’t simply a theoretical exercise; Pasqal deployed its quantum processor to tackle a computationally intensive task with real-world relevance, underpinning applications ranging from energy grid optimization to financial modeling. The core of this advancement lies in the implementation of logical qubits, which are designed to encode information across multiple physical qubits, thereby creating more stable and error-resistant units. Researchers systematically compared the performance of quantum kernel algorithms running on both physical and logical qubits, revealing an average accuracy improvement exceeding 50% with the logical qubit approach.
On a particularly challenging nonlinear problem, the improvement reached a factor of ten. Running on Pasqal’s neutral-atom processor, which has achieved a gate fidelity of 99.4%, the team purposefully pursued logical qubit computation once hardware performance reached a level where application-grade results were achievable. The selection of differential equations as a benchmark is also significant; these equations are ubiquitous in scientific and industrial modeling, and their accurate solution is often computationally demanding.
More importantly, by running a complete application rather than isolated subroutines, we were able to identify precisely which error sources matter most for this class of problem. That understanding is what will guide the next phase of hardware development.
Loïc Henriet, Chief Technology Officer at Pasqal
PROQCIMA Program Drives Application-Driven Research in France
The pursuit of practical quantum computation in France is gaining momentum, driven in part by the PROQCIMA program and yielding results that extend beyond theoretical gains. This achievement isn’t simply a matter of improved accuracy, but a validation of a crucial approach to building reliable quantum computers capable of tackling real-world problems. Pasqal researchers benchmarked their logical qubit approach against conventional methods using a quantum kernel algorithm to solve differential equations, a mathematical cornerstone for modeling phenomena across diverse fields including energy systems, materials science, and financial modeling. The team waited until their hardware reached a combined gate fidelity of 99.4% before attempting application-level computations, recognizing that meaningful results require a foundation of stable, reliable qubits.
This strategic progression is directly linked to the PROQCIMA program, a French initiative designed to foster collaboration between academia and industry in the pursuit of error correction and logical qubit architectures. Henriet explained that this result is a direct product of the PROQCIMA programme, which has created the right conditions for this kind of foundational, application-driven research to happen at scale in France. Pasqal’s ongoing research focuses on further improving gate performance, scaling the number of logical qubits, and refining error detection and correction capabilities, solidifying France’s position in the development of practical quantum computing.
What surprised us is that our logical qubits turned out to be naturally resistant to exactly the types of noise that make solving differential equations harder. We got better results than we had initially anticipated.
