We explore the use of AI agents for solving computational problems in algebra. For the problems we consider, the main challenge is not correctness but time: general-purpose algorithms are guaranteed to provide correct solutions but can be prohibitively slow, while reformulations that exploit special structure often allow solving a problem much faster. We propose to use this setting to evaluate how effectively agents discover and exploit structure within and across tasks to reduce computational costs. More broadly, computational algebra supplies infinitely many problems with independently verifiable answers and systematically controllable structure and difficulty. To illustrate these possibilities, we use benchmark families of systems with controlled shared structure and evaluate agents in three settings: solving independent instances, solving online streams of related instances using persistent memory, and compiling a fast reusable solver for a known problem family. Our experiments show that (i) agents can extend the range of systems solvable within a fixed budget, although reasoning overhead can sometimes outweigh computational savings; (ii) memory can reduce the cost of repeated solving but may preserve suboptimal strategies; (iii) compilation produces faster reusable solvers but at substantial upfront cost. Overall, computational algebra offers concrete, measurable targets for improving how agents discover and reuse structure.
Figures & tables
Figure 1: Hidden-monomial example: the agent reduces degree-171 equations to three cubics and returns all 27 solutions, while direct msolve times out. Agent output is paraphrased.
Direct msolve
Oracle
Agent
Family
Size
solved
s
solved
s
solved
s
Hidden monomial
k=2
12/12
2.2
12/12
0.046
12/12
229
k=3
12/12
228
12/12
0.046
12/12
261
k=4
0/12
600†
12/12
0.046
12/12
230
Hidden triangular
n=12
12/12
9.3
12/12
0.035
12/12
127
n=16
0/12
600†
12/12
0.049
9/12
297
Table 1: Independent-instance setting: solved counts and median wall time per attempt (s) over 12 systems per configuration with a 600 s budget. For eigenvalue systems, m is the matrix size and d the polynomial degree. Direct : msolve on the input equations. Oracle : structural solver using the supplied hidden structure. Agent : Claude Opus 5 without structural hints. Times include unsuccessful attempts and exclude independent verification. † All attempts timed out.
Figure 2: Online streams of 20 quintic systems: (4,1) with coefficient-map degree 6 (left), and (3,1,1) with degree 12 (right). Agent curves show mean cumulative solve-plus-update time over three repetitions on identical inputs and order; shading is ±1 sample standard deviation. Costs include unsuccessful attempts but exclude independent verification. Completion counts are in Table 3 .
Figure 3: Compilation on the (4,1) family. (a) Mean (bold) and individual (light) execution times across three runs, against cumulative compilation time; numbers denote accepted iterations. (b) Projected compilation-plus-execution time after two or five iterations, using mean compilation costs and execution times across runs.
Appendix figures & tables2 assets
Supplementary material from the paper’s appendix.
Appendix
Family
Setting
Public v/e
Degree
Reduced v/e
Solutions
Hidden monomial
k=2
3/3
51
3/3
27
k=3
3/3
102
3/3
27
k=4
3/3
171
3/3
27
Hidden graph
n=12
12/12
9
1/1
9
n=16
16/16
9
1/1
9
n=20
20/20
9
1/1
9
Appendix
Table 2: Independent-instance configurations, with 12 fresh systems each. Degree is the maximum total degree of a public equation; v/e gives variables/equations, and Reduced refers to the supplied structural reformulation used by Oracle. Counts are the generic solution counts used in qualification.
Method
(4,1),d=6
(3,1,1),d=12
(2,2,2),d=6
Direct
121.82(20/20)
92.58(20/20)
115.41(20/20)
Oracle
5.40(20/20)
1.12(20/20)
15.10(20/20)
No memory
169.79±5.30(49/60)
121.37±10.68(58/60)
175.21±1.16(51/60)
Planted memory
43.15±13.21(58/60)
37.07±7.68(59/60)
57.80±10.91(57/60)
Learned memory
76.00±8.34(59/60)
42.24±5.85(59/60)
147.03±9.94(57/60)
Appendix
Table 3: Online-stream completion and total cost. Entries are mean minutes ± one sample standard deviation, followed by complete attempts in parentheses. Agent counts aggregate three 20-system repetitions; baseline counts are from one evaluation. Failed attempts and memory updates are included; independent verification is excluded.
Recent advances in AI for Mathematics have focused largely on autoformalization and theorem proving, leaving the role of Computer Algebra Systems (CAS) in agentic LLM workflows underexplored. We propose a ReAct-style agentic setup that combines LLM reasoning with verifiable feedback from SageMath, together with Context7 for the up-to-date documentation. We evaluate this agentic setup across frontier models for solving research-level mathematical problems from the RealMath benchmark in a setting that emulates a computational-mathematics research loop. We also propose a refinement to the RealMath benchmark by introducing a multi-step post-processing procedure and a multi-stage validation pipeline, both of which improve the quality and reliability of the extracted problem set. Our experiments reveal substantial performance gains from SageMath access across all evaluated models on +9.7pp on average, the gains range from 1.5pp to 27.8pp and narrow the gap between open-weight and closed models. Qwen3.7-Max benefits from SageMath the most, while GPT-5.5 achieves the highest solve rate of 75.2% and the lowest token usage among tool-enabled configurations. Our findings suggest that CAS-augmented agents represent a promising direction for assisting mathematicians in computational exploration, and we believe that this work is a step towards automated conjecture discovery. The project repository is available online.
Pavel Snopov, German Magai
School of Mathematical and Statistical Sciences, The University of Texas Rio Grande Valley, USA · Noeon Research, Tokyo, Japan.
Recent advances in large language models and agentic AI systems have enabled significant progress in mathematical discovery, from solving competition problems to tackling research-level conjectures. However, open problems in computational mathematics have received comparatively less attention: research in this area often requires not only proofs but also numerical experimentation, adversarial constructions, and algorithm design. In this paper, we introduce an agentic research system, Iteris, designed for open problems in computational mathematics. We apply Iteris to two open problems from a recent Simons Workshop collection (arXiv:2602.05394). In these case studies, Iteris generated numerical evidence, constructions, and proof drafts that led, after expert review and correction, to verified results. The first result is a phase diagram for the asymptotic comparison between conjugate gradient and randomized coordinate descent on power-law spectra; the second is a counterexample showing that QR factorization with column pivoting can fail to select well-conditioned submatrices even under low coherence. These case studies suggest that agentic AI systems can participate meaningfully in research workflows for open problems in computational mathematics, while human validation remains essential.
Leheng Chen, Zihao Liu, Wanyi He +1
School of Mathematical Sciences, Peking University · 5Zhongguancun Academy · 2Beijing International Center for Mathematical Research and the New Cornerstone Science Laboratory, Peking University +2
We introduce qtBench, a benchmark of 29 open problems in algebraic combinatorics. These problems ask for a simple combinatorial algorithm that explains given polynomial coefficients or identities. To help AI agents search for such algorithms, we provide a scaffold for playing with combinatorial objects, countering cheating attempts, and testing candidate solutions. To explore how far frontier AI systems can go on these problems, we collaborated with OpenAI, whose researchers obtained proposed solutions to several of them. This highlights bigger meta-problems: is AI discovery at scale ethical? How should these results be released? How should proof digestion be organized? We discuss these questions and give our views.
Eugenio Cainelli, Lorenzo Luccioli, Alessandro Iraci +1