ROOT / COMPUTE // QUANTUM COMPUTERS
// Compute · decoded · publication-grade pass
Quantum Computers, decoded.
The qubit race, judged on what matters in 2026 — error correction, not raw count.
CURATED INDEX◊ VERIFIED · multi-source
120
Nighthawk (advantage push)
0.01%
lowest error (Quantinuum)
2029
IBM fault-tolerant target
// the signal · what the data says
The 2026 quantum story shifted from qubit count to error correction. IBM Nighthawk (120 qubits, tileable to 360 across three modules) targets verifiable quantum advantage by end-2026 using IBM's qLDPC code — far more qubit-efficient than surface codes — on the road to the fault-tolerant Starling system in 2029. Google Willow (105 qubits) demonstrated below-threshold error correction and a verified quantum speed-up on real hardware. Trapped-ion machines (Quantinuum, IonQ) still lead gate fidelity; IBM's Condor holds the raw-count record but count alone no longer wins.
// comparison · every entry, benchmarked & sourced
| Name | Maker | Qubits | Error (%) | Coherence (us) | Modality |
|---|
| Nighthawk | IBM | 120 | 0.3 | 150 | Superconducting |
| ◊ Quantum-advantage target end-2026; qLDPC; 360 via 3 modules |
| Willow | Google | 105 | 0.1 | 100 | Superconducting |
| ◊ Below-threshold error correction; verified speed-up |
| Helios | Quantinuum | 96 | 0.01 | 15000 | Trapped ion |
| ◊ Best gate fidelity; logical-qubit demos |
| Condor | IBM | 1121 | 0.5 | 120 | Superconducting |
| ◊ Largest physical count (2023), now a stepping stone |
| Forte / Tempo | IonQ | 36 | 0.02 | 10000 | Trapped ion |
| ◊ High fidelity, low count, networked |
| Aquila | QuEra | 256 | 1 | 5 | Neutral atom |
| ◊ Analog neutral-atom array |
// leaderboard · Qubits
01Condor1121
02Aquila256
03Nighthawk120
04Willow105
05Helios96
06Forte / Tempo36
// sources & where to verify
// full index, per-field sourcing & CSV export [ AUTHENTICATE → ]
◊ Publication-grade pass, mid-2026. Cross-checked against the sources listed. Green = best value in column. Footnotes: Error is representative 2-qubit gate error. Coherence is typical T2. The field's real 2026 metric is logical (error-corrected) qubits, which physical counts do not capture.