Claude-Designed Quantum Circuits Cut Resource Needs Up to 39 Times in Harvard, Anthropic and QuEra Study

Researchers from Harvard University, Anthropic and QuEra Computing used Claude to design quantum circuits that need fewer resources, QuEra announced. The best designs cut a combined gate and qubit metric by up to 39 times.

The preprint looked at eight building blocks used in larger quantum algorithms. Claude proposed circuit designs, and an independent automated evaluator checked each one for correctness and scored its Toffoli gate count and qubit count. In seven of the eight cases, the AI-designed circuits needed fewer gates and fewer qubits than the baseline designs.

Claude also helped find a new reversible approach to modular inversion, an operation central to elliptic curve calculations. In a complete point addition benchmark, one design needed about 17 times fewer gates than the prior design with the fewest qubits, while using about 12% more qubits.

The combined metric multiplies the Toffoli gate count by the peak number of logical qubits, so a 39 times improvement is not a 39 times speedup. The results are resource estimates for individual building blocks, not runs on quantum hardware.