Infleqtion said on September 24, 2026, that it had entangled 30 logical qubits on Sqale, its neutral-atom quantum computer, using 80 physical qubits. In a technical post announcing the result at Quantum World Congress in College Park, Maryland, Chief Technology Officer Pranav Gokhale wrote that the team completed the experiments in August.
In a neutral-atom machine, each physical qubit is a single atom held in place by a focused laser beam. A logical qubit is a unit of quantum information encoded across several physical qubits so that errors can be detected or corrected.
Infleqtion can detect any single error in its 30 logical qubits but has no guarantee of correcting one, apart from the loss of an atom.
The code
Gokhale wrote that the team grouped the 80 atoms into 10 blocks of eight, each encoding three logical qubits. Each block was prepared in the [[8,3,3]] code and then operated in the [[8,3,2]] code. In that notation, the three numbers are the count of atoms in a block, the number of logical qubits they encode, and the code’s distance.
A code’s distance is the smallest number of physical errors that can change the stored information without being detected. A distance-2 code is guaranteed to detect any single error but not to correct one. Guaranteed correction of any single error requires a distance of at least 3.
Gokhale wrote that the [[8,3,2]] code “can detect any error or correct for a lost atom.” Gokhale wrote that when an atom is lost, the team can reconstruct its missing measurement from parity constraints, fixed relationships among the measurement results within a block. The company calls this technique loss correction.
The circuit and the benchmark
The test was an instantaneous quantum polynomial (IQP) circuit, a type of quantum program whose main layer of basic operations, called gates, can be applied in any order. The team used entangling operations, including four logical doubly controlled Z (CCZ) gates, to connect all 30 logical qubits.
A CCZ gate acts on three qubits at once and is not part of the Clifford set, a restricted class of operations that ordinary computers can simulate efficiently. Infleqtion said the circuit contained 136 physical two-qubit gates and about 1,000 physical operations in total, an amount the company called one KiloQuOp.
Of the more than one billion possible 30-bit outputs, an error-free run of the circuit can produce 262,144, or one in 4,096. Infleqtion reported that about 25% of its measured outputs were in that set, roughly 1,000 times the rate expected from uniformly random outputs.
In its press release, the company described the result as “experimentally confirmed by a signal that is approximately 1000x stronger than underlying noise.” In his post, Gokhale called that 25% share the hit fraction and put it at roughly 1,000 times the random baseline. Infleqtion said that by applying loss correction it quadrupled the number of good shots, runs with outputs among the 262,144, at the cost of higher error rates.
Unpublished details
Gokhale gave the code’s distance in his post, but Infleqtion did not mention it in its press release. Infleqtion did not say in either document what share of runs the team discarded after its error checks, a practice called post-selection. Gokhale wrote that Infleqtion plans to release a full paper with experimental methods, resource counts, and statistical analysis “in the coming weeks.”
The AI-assisted gate
Infleqtion said its researchers found a new logical entangling operation, which they call a double-CZ, with help from OpenAI’s GPT-5.6 Sol model. Gokhale wrote that the model helped the team “discover a novel way to entangle across triplets of logical qubits” and that the double-CZ halved the cost of that entanglement to four physical two-qubit gates. He wrote that, to the team’s knowledge, the only mechanism previously known for it was a transversal controlled-NOT, which uses eight physical gates applied between matching atoms of two blocks.
Gokhale credited a 2023 Nature paper by a Harvard-led team for the transversal CCZ gate in the [[8,3,2]] code, a gate applied atom by atom across a block.
Company and roadmap
In preprints, papers posted before peer review, Infleqtion reported two logical qubits in December 2024 and 12 in September 2025. In a release that month, the company said the 12 logical qubits, achieved “with error detection and loss correction,” put it ahead of schedule for “its previous target of achieving 10 logical qubits with error correction in 2026.” In the same release, it set a target of 30 logical qubits in 2026. The company said it was now “the first neutral-atom quantum computing company to reach 30 logical qubits on a commercial system.”
Chief Executive Matt Kinsella credited co-design of hardware and software with making the demonstration possible on 80 physical qubits. Infleqtion said it has three customers for logical-qubit circuits, among them the Wellcome Leap Quantum for Bio program. Infleqtion’s shares began trading on the New York Stock Exchange under the ticker INFQ in February, after the company completed a business combination with Churchill Capital Corp X, a special-purpose acquisition company.
Gokhale wrote that the team’s priorities include real-time control and mid-circuit measurement “to go beyond post-selection.” With those capabilities, the machine could find and correct errors during a computation. Infleqtion said it remains on track for 100 logical qubits in 2028 and 1,000 by 2030.
