TL;DR
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
Researchers from Anthropic and Harvard used Claude to design quantum circuits for eight computational tasks. Their preprint reports resource-estimate improvements in seven designs, including a decoder requiring about 440,000 Toffoli gates instead of 5.1 million; none of the circuits was demonstrated on quantum hardware.
Researchers from Anthropic and Harvard report that they used Claude to design quantum circuits with lower estimated resource needs across eight computational tasks. In one example, a redesigned decoder used about 440,000 Toffoli gates, compared with 5.1 million in the study’s baseline, but the work was evaluated through computational checks and resource estimates—not tested on a working quantum computer.
The team’s preprint, posted to arXiv, describes circuits for building blocks used in chemistry, optimization, physics simulation and number theory. According to the researchers, seven of the eight designs reduced both the number of costly quantum operations and the number of qubits relative to the study’s comparison designs. The eighth introduced an arithmetic design with different trade-offs between gate count and qubit use.
For the highlighted decoder, the estimate fell from 5.1 million Toffoli gates to roughly 440,000. Its estimated peak qubit requirement also declined, from 1,938 to 1,837. Toffoli gates are a resource measure emphasized in the paper because they can be costly in proposed error-corrected quantum computers. These figures describe the circuit designs and the study’s estimates; they do not show a completed calculation on quantum hardware.
The researchers used an AI-assisted search process rather than relying only on the model to suggest circuit ideas. Anthropic’s accompanying blog post says the team supplied programming tools and an evaluator to check candidate circuits and score resource use, allowing Claude to revise its designs based on feedback. The post identifies the model as Claude Mythos 5 and says its elliptic-curve arithmetic work took 53 hours. Those details are Anthropic’s account of the process.
Lower Costs for Quantum Algorithms
Quantum algorithms are built from smaller routines, and the resources those routines require can determine whether a larger calculation is feasible on a future machine. Reducing gate counts or qubit requirements could, if the designs work as intended when integrated, ease some of the hardware demands of running such algorithms.
The reported decoder result is a large reduction in one estimated operation count, but it is not evidence that a useful quantum application now runs faster or on smaller hardware. Practical gains depend on how the circuit performs as part of a complete algorithm, how it is compiled for a particular machine, and the error-correction and hardware assumptions used. The study points to a possible role for AI in exploring design choices that are difficult to search by hand, while leaving those implementation questions open.
quantum computing circuit design tools
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
How Reversible Circuits Shape Costs
A quantum circuit is a sequence of operations on qubits, the units used to store quantum information. In many quantum algorithms, intermediate results cannot simply be overwritten as in ordinary software: the computation must preserve enough information to reverse operations. Designers often need to clear temporary work through uncomputation, which can add gates, or keep it in memory, which consumes qubits.
The paper describes the model searching across these competing costs, including ways to recover information when needed, share arithmetic across computational paths, and reduce resources devoted to rare inputs. In the decoder example, the team replaced an approach based on the extended Euclidean algorithm with Berlekamp–Massey. The researchers say the design also limited storage for exceptional cases, accepting a small probability of failure under the conditions described in the study.
The decoder is part of decoded quantum interferometry, an algorithmic approach to optimization; it is not, in this use, a circuit for correcting faults in quantum hardware. That distinction matters because the result concerns a routine within a proposed computation, not a demonstrated advance in machine-level error correction.
“The model used was Claude Mythos 5.”
— Anthropic’s accompanying blog post
As an affiliate, we earn on qualifying purchases.
Hardware Tests Still Lie Ahead
The reported results come from computational checks and resource estimates, not runs on a quantum computer. The source material does not establish how the circuits would perform on a specific machine, or whether the estimated savings would hold after integration into complete algorithms and hardware workflows.
The comparison depends on the baselines and assumptions used in the preprint. The supplied report does not provide enough detail to assess every task’s full resource figures or independently compare the designs. For the decoder, the researchers describe a trade-off involving rare inputs and reduced storage; the practical effect of such a trade-off depends on the target algorithm’s requirements and acceptable failure probability.
As an affiliate, we earn on qualifying purchases.
From Estimates to Full Algorithms
The next test is whether the designs can be incorporated into full quantum algorithms and retain their estimated benefits under realistic implementation constraints. That would require further analysis of circuit correctness, resource costs and how the routines map to target hardware; a hardware demonstration is not reported in the supplied material.
The work is a preprint, so its claims should be read as research findings awaiting further scrutiny, not as a validated performance result from a quantum machine. The source material does not specify a publication or hardware-testing schedule. The central open question is whether the reductions in individual building blocks translate into meaningful savings for end-to-end calculations.
quantum computing hardware components
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
What did the Anthropic and Harvard researchers report?
They reported using Claude to design circuits for eight computational tasks. The preprint says seven designs reduced both costly operation counts and qubit requirements against the study’s baselines; the eighth offered a different arithmetic trade-off.
How much smaller was the highlighted decoder?
The researchers estimated about 440,000 Toffoli gates for the redesigned decoder, compared with 5.1 million in the comparison design. Its estimated peak qubit count was 1,837 rather than 1,938.
Did the researchers run the circuits on a quantum computer?
No such demonstration is reported. The results were assessed through computational checks and resource estimates, so performance on working quantum hardware remains unconfirmed.
Which Claude model did the team use?
Anthropic’s accompanying blog post identifies the model as Claude Mythos 5. That identification comes from the company’s account of the work.
What could determine whether the savings matter in practice?
The designs would need to be integrated into complete algorithms and evaluated against the constraints of real machines. The reported reductions alone do not establish that an end-to-end quantum calculation will use less hardware or run faster.
Source: rss
Fall Picks
fall essentials
As an affiliate, we earn on qualifying purchases.
