Search will be available as the English edition grows.
Quantinuum reports 99.9995% logical memory fidelity on its Helios system. The result shows strong error suppression with a color code, while scaling to useful fault-tolerant computation remains unproven.
Quantum computers are vulnerable to noise. A central strategy is therefore to combine multiple physical qubits into a logical qubit and use quantum error correction to detect and correct faults.
What matters is not simply whether error correction works, but whether it becomes more effective as protection is increased. If adding physical qubits introduces problems as quickly as it removes them, the system cannot reach computations long enough to be useful.
Quantinuum now reports a result on Helios that moves in the desired direction: the logical error rate dropped sharply as the researchers increased the strength of the error-correcting code.
The experiment used a color code on Quantinuum's trapped-ion Helios machine.
At the strongest tested code, the company reports logical memory fidelity of 99.9995%. That corresponds to roughly five uncorrectable errors per million error-correction cycles for the measured logical memory.
Quantinuum says this is a tenfold lower error rate than its own Helix demonstration in September. Against the best published result on superconducting hardware cited by the company, Quantinuum reports an improvement of more than 100 times.
The record comparisons are company claims, but the result also drew support from an outside expert. Robin Blume-Kohout of Sandia National Laboratories, who was not involved in the work, describes the error-correction performance gains in Quantinuum's announcement as real and significant.
More important than the headline record is how the system behaved as the researchers increased code distance, meaning the strength of protection around the logical information.
Quantinuum tested color codes at distances 3, 5 and 7. With each step, the company reports that the remaining logical memory errors fell by an average factor of 12. Logical gate errors fell by an average factor of 9.
That kind of error suppression is one of the properties a scalable fault-tolerant architecture needs. The goal is for additional error correction to make logical information progressively more reliable instead of hitting a floor where extra qubits stop helping.
The result also shows why there is no single record that determines which error-correcting code is best.
Quantinuum's earlier Helix demonstration used ten physical qubits per logical qubit. The color-code experiment used up to 37. In return, the color code achieved lower logical error rates.
That illustrates an important design tradeoff for future machines: some codes can prioritize lower qubit overhead, while others buy higher accuracy with more hardware. A practical architecture may need to combine different approaches depending on the operation being performed.
Quantinuum plans a universal fault-tolerant quantum computer called Apollo and is targeting 2029. The new result makes that roadmap more concrete, but it does not prove that Apollo will arrive on that schedule.
The experiment demonstrates very low error rates for specific logical memory and gate operations on today's Helios machine. A useful fault-tolerant quantum computer must also scale to much larger computations, integrate more kinds of logical operations and preserve low error rates through long programs.
Blume-Kohout says a scaled-up version of the system could reach the required reliability with fewer than one million physical qubits. That too is an assessment of scaling potential, not a demonstration of such a machine.
The result is therefore best understood as reducing a central technical risk: Quantinuum has shown that stronger protection on Helios can drive logical errors down rapidly. The next challenge is making that behavior survive as the system and its computations grow.
KEEP EXPLORING
Google has demonstrated a new way to produce a critical resource for error-corrected quantum computers. On its Willow processor, the method reduced errors by about 40 times and achieved a reported fidelity of 99.99 percent.
Measurements in an electron-doped copper oxide directly link superconductivity to antiferromagnetic spin fluctuations. The result resolves an important part of the puzzle, but not the mechanism behind the highest transition temperatures.
E(3)-Pose learns the physical symmetries of the fetal brain and estimates head position from rapid 3D MRI. The goal is to let future scanners adjust diagnostic images as the fetus moves.