Designing more efficient quantum circuits with Claude

Why it matters

Anthropic reports that Claude, given plain-language specifications and an automated evaluator, autonomously designed quantum circuits that improve on the best published versions across eight computational primitives spanning chemistry, optimization, physics simulation, number theory, and numerical computing, in some cases by an order of magnitude or more. The work was led by Andi Gu as part of Anthropic's STEM Fellows Program, mentored by Anthropic quantum physicist Alek Kemeny, with Harvard University collaborators Susanne Yelin and Mikhail Lukin, who is also affiliated with QuEra Computing. The context is that today's quantum hardware remains scarce and error-prone, so reducing the qubits and gates that algorithms require is a key route to practical advantage. Anthropic notes that past reductions in resource requirements came largely from redesigning circuits rather than improving hardware. The reported gains came from Claude uncovering structural and mathematical insights, such as exploiting molecular symmetries to reuse calculations and removing costly components needed only for rare inputs, while working in a closed loop with no human input between rounds.