IBM's quantum results are finally out of reach of classical computers, and you can verify them

IBM's Quantum Results Are Finally Beyond Classical Computers, and You Can Verify Them

Quantum computing has an awkward confession buried in most quantum advantage claims. When a quantum computer solves a problem that no classical machine can, there is often no reliable way to confirm the quantum computer got the right answer. The result sits beyond the reach of the very tools you would use to audit it. That gap has shadowed the field for years, and it is exactly the gap IBM says its three newest quantum advantage results begin to close. 1 2

On July 30, 2026, IBM added three new entries to its quantum advantage tracker, each built around a different strategy for verifying results that classical computers cannot reproduce. 2 The headline claim is not that quantum hardware did something useful. None of the three results has an immediate practical application. 1 The point is that each team found a way to build trust into the computation itself, rather than asking anyone to take the output on faith.

"Trusted computing when you can do classical simulations is irrelevant," Jay Gambetta of IBM told Ars Technica. "Trusted computing when you can't do classical simulations is a big deal." 1

That distinction matters for anyone building systems that might one day depend on quantum results. If you are a bank eyeing quantum optimization for portfolio risk, or a pharma company exploring quantum drug simulation, the question is not just whether the quantum computer is faster. It is whether you can trust the answer when no classical machine can serve as a referee. IBM's tracker exists in part because several past quantum advantage claims saw their lead shrink or vanish once algorithm developers found better classical approaches. 1

The most consequential of the three results comes from a collaboration among Qedma, which develops software to help mitigate errors in quantum processors, RIKEN in Japan, and BlueQubit. 1 2 They modeled Floquet dynamics, the behavior of a quantum system subjected to repeated energy pulses, using circuits of up to 74 qubits. 2 Two state-of-the-art classical simulation algorithms, run on one of the world's largest supercomputers at RIKEN, disagreed with each other in the most demanding regime. 2 One showed magnetism decreasing smoothly; the other showed it increasing. 1 Meanwhile, the quantum computation kept resolving the dynamics. 2

Then the team did something unusual: they ran the same computation on a Quantinuum processor, hardware built by a separate company, and got consistent results. 1 2 That cross-check on different hardware is the closest thing quantum computing has to an independent audit. If the same physics shows up on machines built by different manufacturers with different architectures, the result is far less likely to be an artifact of one company's hardware.

A second result, from IBM and the University of Chicago, attacks the verification problem from a different angle. They built circuits using mostly Clifford gates, which classical computers can simulate efficiently, then strategically inserted non-Clifford T gates that make the overall circuit exponentially hard to simulate classically. 2 "It's got a stronger complexity argument because of the T gates, which you can prove on average is exponentially hard to sample for a classical computer," Gambetta told Ars Technica. 1 The team embedded the computation in a spacetime code, distributing extra qubits across both the circuit's width and its timeline to detect errors during the computation and discard runs where errors occurred. 2 The approach achieved roughly a 10x reduction in effective gate error in a 70-logical-qubit computation. 2 Crucially, the computation carries its own error-detection certificate rather than depending on an external proxy metric afterward. 2

The third entry, from quantum software developer Algorithmiq, took yet another path. Their algorithm estimates the operator Loschmidt echo, a quantity that tracks how information spreads through quantum systems, using 56 qubits. 2 Ars Technica notes the approach resembles Google's quantum echoes work. 1 Algorithmiq's contribution is a strategy for validating the computation process itself, rather than the final answer, which is the only option when no classical machine can verify the answer directly. 2

Taken together, these three approaches sketch a taxonomy of trust for quantum computing. You can cross-check on rival hardware. You can design circuits whose structure makes errors detectable in real time. Or you can validate the process when the answer is fundamentally unverifiable classically. None of these results will optimize a supply chain or break a cipher tomorrow. What they do is move the field from "we are faster, trust us" toward something any auditor could accept: "we are faster, and here is how you can check."

IBM says it expects the back-and-forth to continue as other groups, including Q-CTRL and Birla Institute of Technology and Science, Pilani, submit their own candidates to the tracker. 2 The tracker's competitive structure is itself part of the verification story. Claims go up, classical algorithms try to catch up, and the advantage either survives scrutiny or it does not. For now, these three results survive, not because they are useful, but because they are the first wave of quantum advantage claims that brought their own receipts.