Models & Research

OpenAI opens a science blog with a physicist's account of using GPT-6 Astra on the Standard Model

UC Santa Barbara physicist Nathaniel Craig says GPT-6 Astra helped turn his idea for a lattice definition of the Standard Model into a detailed proof that still needs expert scrutiny.

Open-plan research lab with whiteboards and laptops, no people, logos or readable text
AI-generated

OpenAI has started a science blog, and its first guest post makes a large claim about its newest model. The company said on X late on Sunday 15 November (US Eastern time) that physicist Nathaniel Craig “shares how he worked with GPT-6 Astra to develop a candidate lattice definition of the Standard Model”.

Who wrote it

Craig leads the particle theory group at the University of California, Santa Barbara, and is a visiting scientist at OpenAI, according to the blog post. He says he proposed the construction in an October preprint prepared with the model’s help.

Background

Earlier stories on this thread.

What happened next

Later stories on this thread.

What the problem is

The Standard Model describes the known elementary particles and three of the four fundamental forces. Physicists usually calculate with it by treating interactions as small corrections, an approach called perturbation theory. Craig writes that the full theory still lacks a nonperturbative definition. One route is to model spacetime as a grid, or lattice, but the weak force treats left-handed and right-handed particles differently, and the most natural lattice rules add unwanted partner particles. Earlier work by other physicists handled much of this, Craig says, but the rule for setting a particular phase across the full Standard Model had been open for more than 25 years.

What the model did

Craig says he had a strategy based on representing the force fields on a series of ever finer grids, but could not prove it worked in every case. He writes that GPT-6 Astra suggested how to carry the fields onto finer grids, showed the comparisons would converge and then led the work on the detailed proof. He credits lattice theorists Maarten Golterman, Andreas Kronfeld and Yigal Shamir with checking the proof and asking probing questions.

The caveats

Craig calls the result a candidate definition and says it is more a starting point than a finished answer. He writes that experts still need to decide whether it has all the properties physicists want, that it is an answer in principle rather than a practical tool for simulations, and that he is still examining a property called reflection positivity.

Our read

Analysis: the post is written by a researcher working with OpenAI and published by the company, so it is not an independent assessment. The useful test will be how other lattice physicists respond to the preprint over the coming months.

This report is not legal, financial, or health advice. See the disclaimer.

Primary sources

  1. OpenAI on X: first post on our new science blog OpenAI on X
  2. A step toward defining the Standard Model on a lattice OpenAI

Tags