OpenAI Says GPT-6 Astra Led Proof for Lattice Standard Model
TL;DR
- Nathaniel Craig posted an October 6 preprint proposing a candidate nonperturbative lattice definition of the Standard Model, built with help from GPT-6 Astra.
- Craig credits Astra with turning his strategy into a construction and leading the detailed proof; the paper ships with a Lean 4 formalization.
- The result is the inaugural guest post on OpenAI's new science blog, launched November 15; three named lattice theorists have reviewed the work privately.
Nathaniel Craig, a particle theorist at UC Santa Barbara and visiting scientist at OpenAI, posted an October 6 arXiv preprint proposing a candidate nonperturbative lattice definition of the Standard Model, and credits OpenAI's GPT-6 Astra model with taking the lead on the detailed proof. OpenAI featured the result as the inaugural guest post on a new science blog launched November 15.
The preprint, A gauge-invariant measure for lattice chiral fermions, attacks an obstacle that has stood for more than 25 years since Martin Lüscher's prescription for simpler theories: how to put the Standard Model, a chiral theory, on a four-dimensional lattice without wrecking its gauge invariance. The abstract claims a construction for overlap fermions "in a class of nonabelian chiral gauge theories that includes the Standard Model," and ships with a Lean 4 formalization.
"I had a strategy for tackling this problem, but had not been able to prove that it worked in all the cases it needed to cover," Craig writes. "Astra helped turn that strategy into a construction and took the lead in developing the detailed proof."
Craig is explicit about what has not been done. He calls the result "far more an answer in principle than an answer in practice," with reflection positivity, a continuous-spacetime definition, and realistic lattice simulations all still open. He names three lattice field theorists, Maarten Golterman, Andreas Kronfeld, and Yigal Shamir, as having reviewed the work; no independent published assessment has yet appeared.
It is the 405th OpenAI story on our tracker in the past 90 days, but almost all the rest has been capex, licensing, or the politics around the company's buildout rather than frontier-model output pulled into a specialist proof.
Originally reported by openai.com
Read the original article →Original headline: OpenAI Launches Science Blog as GPT-6 Astra Leads Proof for Candidate Lattice Definition of the Standard Model