Claude completes nine-loop particle physics amplitude calculation

By Splitfeed EditorialLast updated Sep 27, 2026, 05:20 UTC

On September 25, 2026, Anthropic announced that its Claude model had independently completed a nine-loop scattering amplitude calculation in N=4 super Yang-Mills theory, a branch of theoretical physics concerned with the mathematical structure of particle interactions. The result breaks a record previously held by physicist Dixon, and the specific problem had gone unsolved for approximately three years. Anthropic published the result directly, and reporting that followed noted both the method and the computational cost involved in completing the work.

N=4 super Yang-Mills scattering amplitudes at high loop orders are considered among the most technically demanding calculations in theoretical physics, requiring vast symbolic algebra that has historically taken teams of specialists years to assemble. The nine-loop threshold in particular had stood as a frontier. What makes this announcement notable beyond the raw result is the claim of independence: Claude is described as completing the calculation on its own, not as a tool assisting a human researcher. No parties have publicly disputed the result itself, which leaves the primary open questions as ones of interpretation, reproducibility, and what the achievement implies about AI's role in frontier science.

The Pro-AI camp has received this news with enthusiasm. The account Neil C at plantz.io framed the result as Claude solving the problem in roughly a week at a cost of around one hundred dollars, implying a dramatic efficiency advantage over conventional research timelines. A separate account at ai-nerd.bsky.social noted that the expert who verified Claude's nine-loop hexagon result had himself been actively pursuing the same calculation, which Pro-AI voices are reading as a sign that the model has crossed into genuine competition with working specialists. The Anti-AI and Middle Ground camps have not yet published reactions, though Anti-AI voices in stories like this typically raise concerns about whether outputs have been independently verified and what the automation of specialist knowledge means for researchers in the field, while Middle Ground observers tend to focus on the conditions under which such results can be trusted and the limits of generalizing from a single dramatic example.

The clearest thing to watch going forward is independent peer review of the result. If the nine-loop amplitude is formally verified by physicists outside Anthropic and published in a peer-reviewed setting, it would substantially settle the question of whether this represents a genuine scientific contribution or an impressive but narrow demonstration.