GPT-5.2 derives a new result in theoretical physics

2026-02-13 · Source: OpenAI News · Field: Science & Research — Physical Sciences & Chemistry, Mathematics & Computational Sciences, Research Methodology & Innovation · Depth: Expert, short

Summary

On February 13, 2026, OpenAI published a preprint demonstrating that GPT-5.2 Pro conjectured and an internal GPT-5.2 model formally proved a new formula in theoretical physics. The research, titled "Single-minus gluon tree amplitudes are nonzero," identifies specific conditions under which a previously assumed zero-amplitude particle interaction for gluons can occur. This work focuses on scattering amplitudes for gluons, particles mediating the strong nuclear force, specifically when one gluon has negative helicity and the remaining $n-1$ gluons have positive helicity. The preprint shows that in a "half-collinear regime" of momentum space, the amplitude does not vanish, challenging standard textbook arguments. GPT-5.2 Pro initially simplified complex human-derived expressions for $n$ up to 6, then identified a pattern to posit a general formula, which a scaffolded GPT-5.2 version subsequently proved.

Key takeaway

For AI Researchers and Theoretical Physicists exploring complex quantum field theories, this work indicates that large language models like GPT-5.2 can not only assist in simplifying mathematical expressions but also independently derive and prove new theoretical physics results. You should consider integrating advanced AI tools into your research workflow to identify novel patterns and validate conjectures, potentially accelerating discovery in areas previously considered intractable or fully understood.

Key insights

AI can discover and formally prove new theoretical physics results, challenging long-held assumptions.

Principles

• Standard physics arguments assume generic conditions.
• AI can simplify complex mathematical expressions.
• AI can identify patterns for generalized formulas.

Method

GPT-5.2 Pro conjectured a formula by simplifying human-derived expressions and spotting patterns. A scaffolded GPT-5.2 then formally proved its validity, verified against established recursion relations and theorems.

In practice

• Explore specific kinematic regimes for non-zero amplitudes.
• Apply AI to simplify complex physics equations.
• Use AI to generalize formulas from base cases.

Topics

• GPT-5.2
• Theoretical Physics
• Gluon Amplitudes
• AI-assisted Scientific Discovery
• Quantum Field Theory

Best for: AI Researcher, AI Scientist, Research Scientist

Related on AIssential

• Extending single-minus amplitudes to gravitons — Single-minus graviton amplitudes, previously thought to vanish, are non-zero under specific half-collinear kinematic conditions.
• ‘It is incredible’: How AI is transforming mathematics — Large language models are demonstrating surprising "original thought" in mathematics, solving complex problems with novel strategies.
• AI just discovered new physics in the fourth state of matter — AI can discover new physical laws and challenge long-standing assumptions in complex systems.
• The Human/AI Frontier: A Conversation with Bogdan Grechuk — Mathematician Bogdan, an Associate Professor at the University of Leicester, envisions using AI in a loop to solve complex Diophantine equations, where he invents methods and AI tests their applicability.
• Meta-designing quantum experiments with language models — A transformer-based language model can generate Python code to discover generalizable quantum experiment designs, offering human-interpretable scientific insights.
• A Three-Layer Framework for AI in Scientific Discovery — True scientific discovery in AI hinges on qualitative reasoning for model formation, transcending mere search or execution.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by OpenAI News.