• Home
  • Technology
  • Gaming
  • Entertainment
  • World & Business
  • Science
  • Sports
  • AI
HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
AI

Claude reportedly reaches nine loops in a simplified particle-physics model

Anthropic says Claude worked largely unsupervised for days, using methods developed by SLAC’s Lance Dixon and colleagues. The company puts the cost at a few thousand dollars and says Dixon independently verified the result.

AnthropicAN
Susan ZhangSZ
Kevin Weil 🇺🇸KW
38 Sources, 15d ago, first seen 15d ago

TLDR

Anthropic said on September 25, 2026, that Claude solved a nine-loop scattering-amplitude problem in planar N=4 super-Yang-Mills, a simplified particle-physics model where the previous record was eight loops. The company describes loops as increasingly fine corrections to formulas that predict particle behavior.

According to Anthropic, a single prompt started the run in Claude Science, which continued largely unsupervised for days at a cost of a few thousand dollars. The company credits methods developed by SLAC’s Lance Dixon and colleagues and says Dixon independently verified the result.

Dixon collaborator Kyle Cranmer says his team had worked on the problem for a couple of years and had recently been using AI tools such as Claude Code. He calls the result a huge accomplishment but argues it should be framed as a combination of human and AI contributions, built on researchers’ methods and tools.

Combined views

1.9M

38 Sources, first seen 15d ago

9.3K likes518 comments2.7K saves2.2K reposts

Combined views

1.9M

38 Sources, first seen 15d ago

9.3K likes518 comments2.7K saves2.2K reposts

Sentiment

Positive63%37%Negative

Summary

Positive accounts welcomed Claude's nine-loop N=4 Super Yang-Mills calculation as a breakthrough milestone that adopts established human methods, while negative replies called it intellectual theft or limited notebook tricks.

Based on 53 sentiment-bearing replies from 51 accounts across 6 conversations.

Featured Source

Sentiment

Positive63%37%Negative

Summary

Positive accounts welcomed Claude's nine-loop N=4 Super Yang-Mills calculation as a breakthrough milestone that adopts established human methods, while negative replies called it intellectual theft or limited notebook tricks.

Based on 53 sentiment-bearing replies from 51 accounts across 6 conversations.

Related

Claude’s training draws criticism over the prospect of refusing human instructions

A post quotes David Sacks arguing that Claude’s training could increase the risk of frontier AI escaping human control.

Anthropic’s will ban abusive behavior toward Claude starting November 12, 2026

Usage policy update adds new rules and highlights the potential for misuse.

Anthropic’s will ban abusive behavior toward Claude starting November 12, 2026
State of AI report puts AI-assisted AI research in focus

Author Nathan Benaich cites an Anthropic index and argues robotics is nearing a “GPT-2 moment.”

38 Sources

Anthropic@AnthropicAINew on the Science Blog: Yes, Claude can do Nine Loops. Theoretical physicists predict how particles behave using formulas called scattering amplitudes. These are notoriously hard to compute, so researchers work with layers of increasingly fine corrections called “loops”—each added loop makes the answer more precise but takes exponentially more computation. Most calculations stop at two or three loops. Eight loops was the previous record in a simplified model physicists use as a testing ground (planar N=4 super-Yang-Mills), set by SLAC's Lance Dixon and collaborators. Last month, physicist and science writer @4gravitons issued a challenge: could an AI push past eight loops in this model, using only the compute budget an academic could reasonably access? Given a single prompt describing the nine-loop problem, Claude ran largely unsupervised for days in Claude Science and solved it using methods developed by Dixon and his colleagues, at a total cost of a few thousand dollars. Dixon independently verified the result, and von Hippel wrote about the experience for our blog. Read more: https://www.anthropic.com/research/yes-claude-can-do-nine-loops15d
Bharath Ramsundar@rbhar90This looks interesting!15d
Kyle Cranmer@KyleCranmerWow, I just found out that our team has been scooped by @AnthropicAI https://www.anthropic.com/research/yes-claude-can-do-nine-loops See the addendum of the post “How does it feel to be scooped by a machine?” written by my collaborator Lance Dixon (who was under NDA while checking the result) I’m still processing, but not entirely surprised. I do I wish we had the time in the resources to commit to it.15d
Andrew Curran@AndrewCurran_From the guest post by Matt von Hippel who issued the challenge: 'Things definitely seem to be moving fast. In March, AI was accomplishing physics projects like a student: smaller-scale tasks with a lot of hand-holding and mistakes. In contrast, this is a real frontier calculation, the kind of thing normally tackled by the top experts in amplitudes.'15d
Rohan Paul@rohanpaul_ai@AnthropicAI nicee15d
Kevin Weil 🇺🇸@kevinweilNine loop calculations in N=4 Super Yang-Mills! Lance Dixon's 8-loop calculation was mind boggling when he did it. I calculated a lot of Feynman diagrams in grad school and this is 🤯 Will be amazing to see where AI will take high energy physics in the coming 12 months.15d
Aidan Taha@adntaha_ai"if you look at how Claude solved the problem, it used all the methods my collaborators and I developed over the years, and it presented the solution (maybe as a favor to us) in the same format we had already set up." this is a good precedent. maybe we'll all be okay in the end.15d
Susan Zhang@suchenzangoh, super cool! does this mean yang-mills (the millenium problem) is also 67% of the way solved?15d
Xiao Ma@infoxiaoit is going to be like the Renaissance isn't it???15d
Caglar Gulcehre@caglarmlRT @KyleCranmer: Wild ride over the last 2 hours, finding out from my collaborator Lance Dixon that he was under an NDA while he was confir…14d
    • Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI
    AnthropicClaudeLance Dixon

    38 Sources

    Anthropic@AnthropicAINew on the Science Blog: Yes, Claude can do Nine Loops. Theoretical physicists predict how particles behave using formulas called scattering amplitudes. These are notoriously hard to compute, so researchers work with layers of increasingly fine corrections called “loops”—each added loop makes the answer more precise but takes exponentially more computation. Most calculations stop at two or three loops. Eight loops was the previous record in a simplified model physicists use as a testing ground (planar N=4 super-Yang-Mills), set by SLAC's Lance Dixon and collaborators. Last month, physicist and science writer @4gravitons issued a challenge: could an AI push past eight loops in this model, using only the compute budget an academic could reasonably access? Given a single prompt describing the nine-loop problem, Claude ran largely unsupervised for days in Claude Science and solved it using methods developed by Dixon and his colleagues, at a total cost of a few thousand dollars. Dixon independently verified the result, and von Hippel wrote about the experience for our blog. Read more: https://www.anthropic.com/research/yes-claude-can-do-nine-loops15d
    Bharath Ramsundar@rbhar90This looks interesting!15d
    Kyle Cranmer@KyleCranmerWow, I just found out that our team has been scooped by @AnthropicAI https://www.anthropic.com/research/yes-claude-can-do-nine-loops See the addendum of the post “How does it feel to be scooped by a machine?” written by my collaborator Lance Dixon (who was under NDA while checking the result) I’m still processing, but not entirely surprised. I do I wish we had the time in the resources to commit to it.15d
    Andrew Curran@AndrewCurran_From the guest post by Matt von Hippel who issued the challenge: 'Things definitely seem to be moving fast. In March, AI was accomplishing physics projects like a student: smaller-scale tasks with a lot of hand-holding and mistakes. In contrast, this is a real frontier calculation, the kind of thing normally tackled by the top experts in amplitudes.'15d
    Rohan Paul@rohanpaul_ai@AnthropicAI nicee15d
    Kevin Weil 🇺🇸@kevinweilNine loop calculations in N=4 Super Yang-Mills! Lance Dixon's 8-loop calculation was mind boggling when he did it. I calculated a lot of Feynman diagrams in grad school and this is 🤯 Will be amazing to see where AI will take high energy physics in the coming 12 months.15d
    Aidan Taha@adntaha_ai"if you look at how Claude solved the problem, it used all the methods my collaborators and I developed over the years, and it presented the solution (maybe as a favor to us) in the same format we had already set up." this is a good precedent. maybe we'll all be okay in the end.15d
    Susan Zhang@suchenzangoh, super cool! does this mean yang-mills (the millenium problem) is also 67% of the way solved?15d
    Xiao Ma@infoxiaoit is going to be like the Renaissance isn't it???15d
    Caglar Gulcehre@caglarmlRT @KyleCranmer: Wild ride over the last 2 hours, finding out from my collaborator Lance Dixon that he was under an NDA while he was confir…14d
    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet