Quantinuum has introduced Helios, its most advanced quantum computer yet, setting new records in performance, fidelity, and scalability.
Building on its predecessor H2, the system nearly doubles the qubit count and moves further into the quantum advantage regime.
With 98 fully connected qubits and the highest reported accuracy in the industry, Helios represents a major leap in commercial quantum computing.
CEO Dr. Rajeeb Hazra described Helios as “a seamless fusion of hardware and software, creating a platform for discovery unlike any other.”
Designed to bring quantum programming closer to the ease of classical computing, Helios is already being used for enterprise-grade applications.
During an early access program, partners such as SoftBank Corp. and JPMorgan Chase conducted commercially relevant research on the system.
Quantinuum also used Helios to simulate high-temperature superconductivity and quantum magnetism, showing how the technology can translate into real-world use.
The system is now available through Quantinuum’s cloud platform and as an on-premise deployment, with optional integration for NVIDIA GB200 chips.
Helios sets a new benchmark for precision in quantum computing. It delivers single-qubit gate fidelity of 99.9975 percent and two-qubit gate fidelity of 99.921 percent, the highest figures reported in any commercial system so far.
Lead architect Dr. Anthony Ransford said, “You would need to harvest every star in the universe to power a classical machine that could do the same calculations.”
In benchmark tests using Random Circuit Sampling (RCS), the same method used by Google to demonstrate quantum supremacy, Helios achieved results that would be practically impossible for any classical supercomputer.
Quantinuum reported that matching its performance would require more power than the Sun produces, while Helios completed the task using roughly the energy of a single data center rack.
Scalable architecture and new ion design
Helios introduces a new ion-trap architecture that switches from ytterbium to barium qubits. The use of visible-light lasers instead of ultraviolet reduces costs and improves reliability.
Barium also enables the system to detect and remove certain types of atomic-level errors, improving computation quality.
The system’s ion trap features a junction-based Quantum Charged Coupled Device (QCCD) layout that routes qubits efficiently. This allows cooling, sorting, and computation to occur in parallel, improving both speed and accuracy.
Chief Scientist Dr. John Gaebler said the design, first drawn as a napkin sketch, was “incredible to see realized after all the team’s work.”
Real-time programming and fault tolerance
Helios introduces a real-time control engine that allows programs to adapt as they run.
It can interleave GPU-accelerated classical and quantum computations within a single program, enabling faster problem-solving and more flexible algorithms.
The system converts 94 of its 98 physical qubits into logical ones, fully entangled in one of the largest GHZ states ever recorded.
These logical qubits, built on Quantinuum’s Iceberg code, achieve near break-even fidelity, meaning they outperform unencoded physical qubits.
Helios also achieved 48 fully error-corrected logical qubits at a 2:1 encoding rate, a result once thought impossible.
Competing systems would require thousands of qubits to reach the same level of fault tolerance.
With Helios, Quantinuum has positioned itself at the forefront of practical quantum computing, combining scientific breakthroughs with a design ready for real-world deployment.