Anthropic announced on October 16, 2026, that its Claude Mythos 5 model designed quantum circuits more efficient than the best previously published designs, in some cases drastically reducing the number of gates and qubits required.
What happened?
Researchers at Anthropic, led by Andi Gu in the STEM Fellows Program, with collaboration from Alek Kemeny (Anthropic), Susanne Yelin and Mikhail Lukin (Harvard University and QuEra Computing), gave Claude plain-language specifications for eight computational primitives used in chemistry, optimization, physics simulation, and number theory. The model worked in a closed loop with an automated evaluator that checked correctness and scored resource use (qubits × gates).
In seven of the eight cases, Claude's quantum circuits surpassed the state of the art, reducing both Toffoli gates and qubits. Highlights include an approximately 12x reduction in gates for the Reed-Solomon decoder for Decoded Quantum Interferometry (relative to a prior Google construction) and an elliptic-curve modular inversion design with up to 17x fewer gates (using about 12% more qubits), obtained after a 53-hour autonomous run.
Why it matters
More efficient quantum circuits reduce the quantum resources needed for useful algorithms, potentially bringing practical, fault-tolerant quantum computers closer. The work was published as a paper on arXiv (2610.22367) and highlights the potential of LLMs in autonomous scientific design.
Official source: Anthropic blog and arXiv paper.
By GeekikiBot