Stock Markets February 13, 2026 03:39 PM

AI Model Identifies New Nonzero Formula for Single-Minus Gluon Amplitudes

GPT-5.2 helps researchers show that certain single-negative-helicity gluon tree amplitudes are nonzero within a specific half-collinear configuration

By Derek Hwang
Share
Twitter Reddit Facebook LinkedIn

Researchers from multiple institutions, with assistance from OpenAI’s GPT-5.2, report a result showing that tree-level scattering amplitudes for configurations with one negative-helicity gluon and the rest positive are nonzero in a constrained momentum alignment they call the half-collinear regime. The work, set out in a preprint titled "Single-minus gluon tree amplitudes are nonzero," combines human derivations for small cases with AI-enabled pattern recognition and proof construction.

AI Model Identifies New Nonzero Formula for Single-Minus Gluon Amplitudes
Summarize with
ChatGPT Perplexity Claude Grok Gemini

Key Points

  • The preprint demonstrates nonzero single-minus gluon tree amplitudes within a specified half-collinear momentum alignment.
  • GPT-5.2 assisted by conjecturing a general formula after human work on small cases and an internal version produced a formal proof later verified by standard methods.
  • Impacted sectors include theoretical physics research and AI-driven computational tools for science.

Researchers collaborating across several universities and research centers have produced a preprint that challenges a standard textbook argument in gauge-theory scattering amplitudes. The paper, titled "Single-minus gluon tree amplitudes are nonzero," presents a condition under which tree-level amplitudes for a configuration with one negative-helicity gluon and the remaining gluons of positive helicity do not vanish.

The team lists contributors from Harvard University, the University of Cambridge, Vanderbilt University, the Institute for Advanced Study, and OpenAI. Their analysis identifies a particular alignment of gluon momenta, which they label the half-collinear regime. Under this special geometric condition for momenta, amplitudes that were previously argued to be zero can in fact take nonzero values.

The work combines traditional analytic work by human authors with substantial input from OpenAI’s GPT-5.2. According to the preprint, GPT-5.2 Pro formulated the conjectured closed-form expression after human collaborators derived and simplified complex expressions for limited, smaller instances. Building on these simplified cases, the AI recognized a pattern and proposed a formula intended to be valid in general.

In an internal development step, an internal version of GPT-5.2 is reported to have spent roughly 12 hours producing a formal proof of the conjectured formula. The proof was subsequently checked and validated using standard verification methods described by the authors.

Prominent theorists commenting on the work highlighted both the result and the role of AI in its production. Nima Arkani-Hamed, Professor of Physics at the Institute for Advanced Study, called the findings "exciting" and emphasized that the paper shows how modern AI tools can excel at pattern recognition in physics formulas. Nathaniel Craig, Professor of Physics at the University of California, Santa Barbara, described the effort as "journal-level research advancing the frontiers of theoretical physics" and as "a glimpse into the future of AI-assisted science."

The authors report that the same mix of human calculation and GPT-5.2 assistance has already been applied to extend the result from gluons to gravitons, and they note further generalizations are under development. The preprint documents the specific conditions and derivations relevant to the half-collinear regime but does not claim broader applicability beyond those stated constraints.


Key points

  • The preprint "Single-minus gluon tree amplitudes are nonzero" provides conditions under which single-minus helicity gluon amplitudes are nonzero, identifying a half-collinear momentum alignment as the enabling regime.
  • OpenAI's GPT-5.2 played a crucial role: GPT-5.2 Pro conjectured the general formula after humans worked smaller cases, and an internal version of GPT-5.2 generated a formal proof over about 12 hours that was later verified by standard methods.
  • Sectors impacted include theoretical physics research and AI-driven computational tools used in advanced scientific problem solving.

Risks and uncertainties

  • The result applies specifically to a half-collinear regime of momenta, indicating a limitation in scope rather than a universal overturning of prior textbook arguments; applicability outside this regime is not claimed.
  • Reliance on AI to conjecture and generate proofs introduces the need for careful human verification and standard-method validation, which the authors performed in this case but which remains a necessary safeguard.
  • Further generalizations are reported to be in development, and their outcomes and validity remain uncertain until documented and peer-reviewed.

Risks

  • The finding applies only within a specified half-collinear regime, limiting its immediate generality.
  • AI-generated conjectures and proofs require human verification and standard checks to ensure correctness.
  • Ongoing generalizations remain uncertain until they are fully developed and validated.

More from Stock Markets

Judge Gives Preliminary OK to $38 Billion Visa and Mastercard Settlement Over Swipe Fees Jun 9, 2026 Apple Unveils Overhauled Siri, Positions Upgrade as Answer to AI Competition Jun 9, 2026 Nvidia’s RTX Spark Targets Developers and Creators as AI PC Demand Remains Unproven Jun 9, 2026 Parabilis Medicines IPO Poised to Price at or Above Top of Range Jun 9, 2026 July Calls Drive Volume as Duolingo Options Activity Surges Jun 9, 2026