Marking a major 2026 milestone on the IBM Quantum Roadmap, IBM just unveiled its fastest quantum processor to date: the IBM Quantum Nighthawk r2. By delivering accurate observable estimation on circuits containing more than 7,500 gates, the new chip successfully blasts past previous capability targets.

The system easily executes over 100,000 circuits per second. This unprecedented speed represents up to a 25x leap in circuit throughput compared to the older IBM Quantum Heron fleet, which previously topped out at roughly 4,000 circuits per second.

While Nighthawk r2 maintains the fundamental scale of its r1 predecessor, it delivers highly targeted improvements in quality, with its most significant advance coming in raw speed. From the first demonstrations of quantum advantage to newly introduced modular architectures for cryogenic systems, this processor represents the culmination of numerous important advances in quantum computing this year.

With IBM Quantum Nighthawk r2 now available on the IBM Quantum Platform, that rapid progress continues. New users can currently create a free account and get started on this cutting-edge hardware today.

Design and physical architecture

Under the hood, the Nighthawk r2 integrates 120 programmable qubits alongside 218 dedicated couplers. It further bolsters this array with 120 independent reset elements, explicitly pairing one dedicated reset element with each programmable qubit.

Altogether, the entire layout features a massive 458 physical quantum elements. From an engineering perspective, these additional couplers and reset elements are virtually indistinguishable from the programmable qubits themselves.

Such a dense and intricate hardware configuration makes this the most complex quantum processor IBM has ever pushed into production. Rather than limiting connections to merely two or three neighbors like previous architectural generations, the upgraded square-lattice architecture links most qubits to four nearest neighbors.

This major structural enhancement enables significantly greater complexity and efficiency in modern circuit design. Achieving this level of interconnectivity required incorporating more controllable quantum elements than users ever program directly.

Accelerating computations with independent, high-speed qubit reset

Capitalizing on this advanced design, engineers tackled a longstanding bottleneck in quantum computing: the idle time spent waiting for qubits to reset between circuit executions. Every single quantum circuit must begin with qubits prepared in a known state and end with a qubit measurement.

Between executions, qubits must return entirely to their ground state before the next shot can begin. To achieve this in previous generations of IBM Quantum processors, including Heron, the hardware relied on a technique powered by dynamic circuits called conditional reset.

Under conditional reset, a processor measures a qubit's state and applies a specific bit flip gate—known as a π-pulse—to flip the qubit back to its ground state only if it is found in the |1⟩ state. Unfortunately, conditional reset remained strictly limited by measurement fidelity.

Furthermore, this older technique proved incapable of resetting qubits that had leaked entirely outside the computational state. To ensure a full and reliable reset, older systems were forced to sit idle for hundreds of microseconds between every single circuit execution.

Nighthawk r2 overcomes this critical limitation by replacing traditional conditional resets and idle time with a highly efficient dissipative reset gadget. Each programmable qubit is linked via a high-dynamic-range tunable coupler to a cold environment capable of drawing it back to its ground state directly on demand.

When activated, this coupler forcefully pulls a qubit's effective T1—the precise measure of how long it holds its energy—from a median of about 200 microseconds down to just roughly 25 nanoseconds. This enables a significantly faster, high-quality reset.

Consequently, idle times between runs plummet to as little as a single microsecond. Slashing this downtime is precisely how the Nighthawk r2 consistently reaches 100,000 circuits per second.

The direct result is a dramatic increase in the total amount of useful computation the system can perform over time. In practice, these massive speed gains are most noticeable in the large-scale, repetitive workloads that are highly common in the research community.

Boosting speed without compromising quality

Even the fastest quantum processor offers little real-world value if that speed comes at the cost of the fidelity required for useful, accurate computations. Nighthawk r2 does not just maintain the high quality of its predecessor; it actively introduces a key improvement in error rates.

Crucially, because this active high-speed reset capability cools the qubit directly to its ground state, it drops initialization errors across the device by around 25x. The underlying mechanism is remarkably simple: cleaner starting states inherently mean more accurate computational results.

Nighthawk r2 achieves these remarkably clean starting states while fully maintaining Heron-class gate fidelity during high-speed operations. Because the active qubit reset process is explicitly neighbor-safe, individual qubits can reset immediately without degrading the performance of nearby components.

This neighbor-safe operation remains absolutely crucial for Nighthawk’s densely packed square-lattice architecture. The end result is a revolutionary processor that doesn’t just perform more quantum computation, but vastly more useful quantum computation.

Quantum advantage and other early results on Nighthawk r2

Putting this hardware speed to the test, Nighthawk r2’s improvements are already enabling exciting demonstrations of its quantum computational capability. Early runs on advantage-candidate circuits have yielded up to 10x faster runtimes with absolutely no loss in accuracy.

Following recent demonstrations of quantum advantage through trusted quantum computation, researchers utilized Nighthawk r2 to perform advanced doped Clifford sampling advantage experiments. Created through a joint collaboration between UChicago and IBM, these experiments demonstrate how quantum codes can successfully execute computations beyond the reach of leading classical simulation methods.

Crucially, these experiments accomplish this while simultaneously providing trust that the mathematical computation was executed correctly. With Nighthawk r2’s increased connectivity relative to the older Heron processors, engineers expect to see even more demonstrations of quantum advantage across a broader range of complex circuits and problem classes.

Looking toward future fault-tolerant applications, Nighthawk r2 has demonstrated highly accurate observable estimation using Probabilistic Error Amplification (PEA) on massive circuits containing 7,500 gates. This achievement firmly establishes the system as a critical testbed for the reliable execution of increasingly complex quantum circuits.

The hardware's benefits also further extend to experiments directly exploring real-world applications. During recently reported neutron-scattering simulations, the processor's massive throughput unlocked an unprecedented 12x speedup.

Thanks to this targeted hardware performance, researchers managed to successfully generate spectra for direct comparison against experimental laboratory data in approximately 60 seconds. Together, these impressive results clearly show how advances in throughput and quality are rapidly translating into practical computational capability. These capabilities are currently enabling demonstrations of quantum advantage, ticking off vital targets on the IBM roadmap, and rapidly accelerating the discovery of novel quantum applications.

Built for quantum error correction

Nighthawk r2’s transformative new reset capability is available not only between separate circuits, but also actively during ongoing circuit execution. This functionality brings enormous efficiency gains for dynamic circuits, an advanced capability that involves performing discrete qubit measurements within the runtime of a single circuit execution.

Many modern quantum error correction protocols and advanced quantum-classical workflows strictly require dynamic circuits to function. By providing reliable, independent reset capabilities across all programmable qubits, Nighthawk r2 serves as a highly powerful platform for conducting specialized research in these critical areas.

The new high-speed reset capability enables the repeated use of auxiliary qubits during complex quantum error detection routines. It acts as the primary backbone for quantum error correction protocols like space-time checks.

This inherent flexibility is strictly essential for advanced experiments that actively introduce logical qubits into sophisticated computing workloads. In many distinct ways, this is a direct reflection of Nighthawk’s evolving role in the larger IBM Quantum Roadmap.

Originally introduced to the public as a platform for exploring and scaling quantum advantage, the architecture has now successfully become a premier testbed for enhanced error-corrected operations. Mid-circuit reset features ultimately make Nighthawk r2 a foundational tool for advanced fault-tolerant quantum computing research.

Run circuits today

As quantum hardware steadily matures, the primary challenge is no longer simply building bigger and better physical processors. Instead, the overriding goal is actively maximizing what researchers can accomplish with the specialized systems that are already commercially available.

With its array of 120 programmable qubits and independent, high-speed qubit reset technology, Nighthawk r2 aims to finally free researchers from previous hardware limitations. Greater overall circuit throughput strictly means the new Nighthawk chip can run substantially more computational workloads within the exact same user allocation.

Current internal estimates suggest that eager users may easily realize much more usable computational capacity with Nighthawk r2. In practice, this directly means they can spend significantly less time managing quantum hardware resources and vastly more time discovering what those resources make possible.

For those eager to test its capabilities, the Nighthawk r2 is currently live and openly available to users on the IBM Quantum Platform. Whether you’re scaling application research, deeply exploring the power of dynamic circuits, or aggressively advancing the latest quantum error correction research, this new system will definitively help you do more quantum computing, significantly faster.

Ultimately, the most important thing about a quantum computer isn’t merely the static number of qubits or physical components; it’s the exact amount of useful computation the system can deliver. As the broader tech industry steadily evolves toward more sophisticated workloads and fully fault-tolerant computation, it is absolutely essential that engineers continue working to seamlessly turn hardware innovation into highly practical capability.

Researchers and developers looking to actively learn more about this hardware can register for IBM's upcoming September 10 webinar, officially titled "Quantum computing performance in practice". Head directly to the IBM Quantum Platform today to clearly see what the groundbreaking IBM Quantum Nighthawk r2 can do for your personal research.