IQM reports entanglement across its full 54-qubit chip
A whole-chip benchmark asks how much hardware can work together. Its readout treatment remains part of the result.
IQM says its Emerald processor has produced a GHZ state across all 54 . Its October 29 announcement reports fidelity of 0.516 after readout error mitigation. A separate 40-qubit result reached 0.515 without that treatment. Both reported point estimates exceed the 0.500 threshold used in the reported test. The company says the improved system performance is available through its cloud platform.[1]
The result addresses an intuitive question about a quantum processor: how much of the chip can participate in one coordinated experiment? Counting the components gives one answer about a machine. Testing a state that spans those components asks a different question about their collective behavior.
The two reported fidelities should not become a league table. They come from different system sizes and different treatment of the measurements. Readout mitigation adjusts estimated measurement results using an error model; the qualification is part of the evidence, even when the result is shortened for a card or shared headline.
IQM's September software documentation describes joint optimization of the selected qubits, preparation tree and gate order. That is useful background for how such tests can be arranged, but it does not independently verify the October measurement or identify the exact routine used in it.[2]
The next useful evidence would establish how repeatable the result is, how its uncertainty compares with the threshold and how much work a user needs to reproduce it. A whole-chip state test is meaningful within that scope. It does not establish that arbitrary programs can use every qubit equally well or that a useful computational advantage has been reached.