A new threshold in automated circuit design
Anthropic's Claude Mythos 5 model delivered a significant efficiency gain in quantum circuit design. In a 53-hour experiment led by researcher Andi Gu through Anthropic's STEM Fellows Program, the model developed a reversible arithmetic algorithm for elliptic-curve calculations. Researchers from Harvard University and QuEra Computing also contributed to the work.
The balance between memory and processing time
Quantum computers operate on units called qubits, and existing machines hold only hundreds to thousands of them. Every operation consumes resources, while storing intermediate results requires additional qubits. Quantum circuits also cannot discard temporary information, since their operations must remain reversible.
Claude's method recovers the branching decisions of an algorithm called divsteps directly from intermediate numerical values. This removes the need to store the full history of branching decisions. The model tested the approach through numerical simulations and formalized its correctness using Lean.
Notable results
- One design uses 12 percent more qubits than the smallest previous design but requires 17 times fewer gates.
- The model optimized eight building blocks across chemistry, physics simulation, optimization, number theory and numerical computing.
- The costliest step of a leading quantum optimization algorithm was reduced by roughly 12 times.
According to Anthropic, these results expand the options for engineers working with limited quantum hardware and suggest automated circuit design could shorten the path to practical quantum computing.
