Explainer Collection

Quantum Computing

Nation (2025)

Seven quantum SDKs face the same benchmark, and only Qiskit clears the full transpilation suite

Benchpress runs 1,066 identical tests across BQSKit, Braket, Cirq, Qiskit, QTS, Staq and Tket on circuits up to 930 qubits, exposing two orders of magnitude in compile time and a wide gap in functional coverage.

1,066: tests run against every SDK, on circuits up to 930 qubits and around a million two-qubit gates
1 of 7: SDKs that pass the full 1,032-test transpilation set (Qiskit). Tket fails 87, BQSKit 200, QTS 19

IBM Quantum (2025)

IBM bets on bicycle codes for the first large-scale fault-tolerant quantum computer, and sets a date

Two new papers and a refreshed roadmap commit IBM to delivering Starling, a 200 logical qubit machine running 100 million gates, by 2029.

2029: target year for IBM Quantum Starling, the first large-scale fault-tolerant quantum computer
200: logical qubits Starling will run, encoded across many more physical qubits

Bravyi (2024)

Quantum memory gets surface-code thresholds, but with a tenth of the qubits

Bravyi and colleagues introduce bivariate bicycle LDPC codes for fault-tolerant quantum memory. In simulations under a standard circuit noise model, the best small code reaches a pseudo-threshold near 0.7% and stores 12 logical qubits with 288 physical qubits, where a comparable surface-code layout would need nearly 3,000.

0.7%: pseudo-threshold for the largest reported BB code under circuit-based noise, close to the surface-code threshold
288: physical qubits needed to store 12 logical qubits with the [[144, 12, 12]] BB code