Search arXiv⌕ Search
Home/Archive/arXiv

arXiv · 2610.16733

TheoPhysEval: Assessing AI agents at autonomous theoretical physics problem solving

A. Ramesh Chandra·Aaghaz Mahajan·Ainesh Sanyal·Mianqi Wang·Shubhadeep Chakraborty·Sudip Chakraborty·Shailesh Chandrasekharan·Rama Govindarajan·Akshit Goyal·Jonathan J. Heckman·Yasir Iqbal·Divya Jaganathan·Saumav Kapoor·Andreas Karch·Pushkar Khandare·Sandeep Krishna·Anupam Kundu·Alok Laddha·Tongyan Lin·Swarnendu Maity·Nilmani Mathur·Shani Nadir Meynet·Alessandro Mininno·Abhijeet Mishra·Vidur Niranjan·Abhishodh Prakash·Pratik Rath·Durbar Roy·Debargha Sarkar·Srikanth Sastry·Gary Shiu·Aninda Sinha·Thomas Van Riet·Praneeth Netrapalli·Suvrat Raju
View original source ↗Read source document ↗Export metadata ↓

Abstract

We present a suite of 50 unpublished research-level theoretical physics problems with verifiable answers. Without AI assistance, a typical problem in our collection could be solved by a world expert in about a day or less but would take considerably longer for experts in adjacent specialties. We test the ability of off-the-shelf AI agents to tackle these problems autonomously, without human hints. When given access to tools available in a realistic setting, including the ability to search the literature and write scripts for numerical and symbolic work, the best-performing agent, GPT-6-Astra harnessed in Codex, obtains the final answers with a near-perfect accuracy of 98% at an average cost under USD 3 per problem. On average, agents show only small changes in accuracy when denied web search and access to background research papers, suggesting significant memorization of the literature. These results suggest that state-of-the-art agents can accurately perform concrete calculations required across theoretical physics research. We comment on the broader implications of these findings and discuss possible explanations for the current lack of agent-led breakthroughs in theoretical physics despite their outstanding problem-solving abilities.

Explore related subjects

hep-thcond-mat.stat-mechcond-mat.str-elcs.AIgr-qc

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

A. Ramesh Chandra, Aaghaz Mahajan, Ainesh Sanyal, Mianqi Wang, Shubhadeep Chakraborty, Sudip Chakraborty, Shailesh Chandrasekharan, Rama Govindarajan, Akshit Goyal, Jonathan J. Heckman, Yasir Iqbal, Divya Jaganathan, Saumav Kapoor, Andreas Karch, Pushkar Khandare, Sandeep Krishna, Anupam Kundu, Alok Laddha, Tongyan Lin, Swarnendu Maity, Nilmani Mathur, Shani Nadir Meynet, Alessandro Mininno, Abhijeet Mishra, Vidur Niranjan, Abhishodh Prakash, Pratik Rath, Durbar Roy, Debargha Sarkar, Srikanth Sastry, Gary Shiu, Aninda Sinha, Thomas Van Riet, Praneeth Netrapalli, Suvrat Raju. 2026-10-09. TheoPhysEval: Assessing AI agents at autonomous theoretical physics problem solving. https://arxiv.org/abs/2610.16733

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

Three Point Amplitudes in Matrix Theory

We compute the three graviton amplitude in the Banks-Fischler-Shenker-Susskind matrix model for M-theory. Even though the three point amplitude is determined by super Poincare invariance in eleven dimensional M-theory, it requires a non-trivial computation in the matrix model. We consider a configuration where all three gravitons carry non-zero longitudinal momentum. To simplify the problem, we compactify one additional dimension and relate the amplitude to a supersymmetric index computation. We find agreement with the expected answer even at finite values of $N$.

hep-th↗
arXiv2026

A Note on Kinematic Flow and Differential Equations for Two-Site One-Loop Graph in FRW Spacetime

In this work, we systematically study the differential systems governing loop-level wavefunction coefficients of conformally-coupled scalar field theory within a general power-law FRW cosmology. By utilizing the twisted cohomology, hyperplane arrangements, and IBP techniques, we derive the canonical differential equations for two-site one-loop bubble and tadpole systems, revealing distinct structural differences. We present new insights into the one-loop tadpole system, uncovering that its integral family can include multiple parent functions due to distinct pairs of relative hyperplanes associated with each function, unlike the single parent function appearing in the one-loop bubble case. Moreover, we demonstrate that the tadpole correlator selectively probes only a subset of the cohomology space, despite the hyperplane arrangement suggesting a higher-dimensional structure. Another novel contribution of this work is the extension of kinematic flow framework to the loop-level scenarios for the first time. Using a graphical approach based on family trees generated by marked tubing graphs, which encode singularity structures, we efficiently construct the differential equations and uncover the hierarchical relationships among the associated master integrals. Additionally, we provide a preliminary discussion on generalization to two-site higher-loop configurations. We propose a general decomposition formula for the canonical form of a two-site diagram with arbitrary loops, breaking it into unshifted and shifted components associated with the fundamental tree-level and bubble-like structures, and establish a block-wise decomposition rule for the matrix $A$ in the corresponding differential system. These advancements provide a unified framework for two-site loop-level correlators and lay the groundwork for future study of more complex multi-site loop systems.

hep-th↗
arXiv2026

On anomaly free 4d $\mathcal{N}$=4 and 6d (2,0) conformal supergravities and UV finiteness of Poincaré supergravities

We review the structure of superconformal anomalies in 4d $\mathcal N$=4 conformal supergravity (CSG) coupled to a number N$_\rm v$ of $ \mathcal N$=4 vector multiplets and 6d (2,0) CSG coupled to N$_{_{\rm T}}$ of (2,0) tensor multiplets. Anomalies cancel if N$_\rm v$=4 and N$_{_{\rm T}}$=26 respectively. If the CSG part of the action is dropped and N$_{\rm v}$=6+ n$_{\rm v}$, the first theory is classically equivalent to the 4d $\mathcal N$=4 Poincaré supergravity (PSG) coupled to n$_{\rm v}$ vector multiplets, while the second one with N$_{_{\rm T}}$=5+ n$_{_{\rm T}}$ is classically equivalent to the 6d (2,0) PSG coupled to n$_{\rm T}$ tensor multiplets. We argue that these facts imply that divergences in the 4d PSG with n$_{\rm v}$ vectors should be proportional to n$_{\rm v}$+2 and similarly in the 6d PSG with n$_{_{\rm T}}$ tensors to n$_{_{\rm T}}$-21. These predictions appear to be consistent with known results of explicit scattering amplitude computations in these 4d and 6d PSG theories.

hep-th↗