Call for Papers

Mathematical reasoning is central to science, engineering, finance, education, and mathematics itself. Since the first MATH-AI workshop, the field has moved from asking whether large language models (LLMs) can solve mathematical problems to asking how AI systems can participate across the full range of mathematical research: proposing conjectures, searching for examples and counterexamples, formalizing arguments, proving theorems, designing algorithms, and collaborating with human researchers.

This year, our workshop focuses more squarely on (but is not limited to) the intersection of agentic AI and mathematical reasoning. Recent progress makes this an especially timely moment: AI systems have achieved super-human results on competition-style and formal mathematical reasoning tasks, autoformalization is connecting natural mathematical language with proof assistants and formal libraries, and AI systems are beginning to guide mathematical discovery in topology, representation theory, combinatorics, matrix multiplication, and geometry. These advances point toward automated mathematical discovery, in which AI helps automate parts of the research loop from conjecture generation to proof search, verification, and communication.

This year, our central question is: How can agentic AI systems advance mathematical research while remaining reliable collaborators for human mathematicians? This theme links two priorities: building agents that can plan, use tools, conjecture, formalize, prove, verify, and learn from feedback; and designing human-AI workflows in which such agents extend mathematical judgment. To address this question, we aim to bring together diverse participants from different backgrounds, institutions, and disciplines into our workshop. Our objective is to foster a lively and constructive dialogue on areas related, but not limited, to the following:

• Humans vs. machines: What are the comparative strengths, limitations, and characteristic failure modes of human mathematicians and AI systems? Which aspects of mathematical reasoning, judgment, and creativity remain distinctively human or machine?
• Building reliable mathematical agents: What architectures, training methods, memory systems, tool interfaces, and feedback mechanisms enable agents to plan over long horizons, recover from errors, and operate reliably in mathematical environments?
• Evaluating mathematical agents: How should we evaluate agents on advanced and open-ended mathematical tasks—beyond answer accuracy and formal correctness—to capture proof quality, robustness, creativity, efficiency, and effective use of tools and feedback?
• Automated mathematical research: Which parts of the mathematical research process can agents conduct autonomously—from problem selection and experimentation to conjecture refinement and verification—and how should human researchers supervise, redirect, and collaborate with them?
• Education: What roles can agentic AI systems play in mathematics education—tutoring, guiding exploration, and providing feedback—especially where expert instruction and educational resources are limited?
• Cross-domain applications: How can mathematical agents enable progress in formal verification, software and hardware design, science, engineering, finance, and other domains that depend on complex mathematical reasoning?

Important Dates

Paper submission opens: July 25, 2026
Paper submission deadline: September 6, 2026 (AoE)
Reviewing deadline: September 20, 2026 (AoE)
Author notification: September 29, 2026 (AoE)
Camera-ready deadline: October 3, 2026 (AoE)
Workshop: December 12 or 13, 2026 (TBA)

Due to NeurIPS paper-import policies and related updates this year, we cannot extend the submission deadline as late into September as in prior years (when we aimed to accommodate concurrent ICLR submissions). We still welcome 4-page submissions and encourage authors to submit preliminary drafts to our workshop.

Submission Requirements

Submissions to MATH-AI 2026 are limited to 4 pages of content in the NeurIPS format, but may contain an unlimited number of pages for references and supplementary materials. The latter may not necessarily be read by the reviewers. We request and recommend that authors rely on the supplementary material only to include minor details (e.g., hyperparameter settings, reproducibility information, etc.) that do not fit in the 4 pages. The review process is double-blind, so please ensure that all papers are appropriately anonymised.

LaTeX template: Download the MATH-AI NeurIPS 2026 template. Submissions should use the NeurIPS 2026 workshop format. Please use \usepackage[dblblindworkshop]{neurips_2026} for submission and \usepackage[dblblindworkshop, final]{neurips_2026} for the camera-ready version.

All accepted papers will be presented in an in-person poster session, and some will be selected for oral presentation. Previously published work (including NeurIPS 2026) is not allowed. Accepted papers will be displayed on the MATH-AI 2026 Papers page, but are to be considered non-archival. We will select one Best Paper Award and up to three Best Paper Honorable Mentions.

Submission site: OpenReview. If you haven’t already, we recommend creating an OpenReview profile as soon as possible, as new profile creation can take up to two weeks in some cases.

Reviewing Process

This workshop uses reciprocal reviewing: for each submission, at least one author must agree to review. Each person may serve as the reciprocal reviewer for at most one submission, and we aim to assign about three papers per reviewer; final assignments depend on expertise and conflicts. We are also calling for additional reviewers: please nominate yourself or others via our reviewer nomination form. We are committed to a rigorous and fair evaluation of all submissions. Reviewers must declare both domain and individual conflicts before review assignments are made. Two to three organizers will serve as publication chairs, making final decisions based on reviews. Chairs will recuse themselves from any submission with which they have a conflict, as identified by OpenReview. Each submission will receive at least three reviews, and there will be no rebuttal phase. Final decisions will rely heavily on reviewer feedback; exceptions will be made only in rare cases and must be approved by at least two non-conflicted organizers. Outstanding reviewers will receive an Outstanding Reviewer Award, and all reviewers will be publicly acknowledged on our Papers page unless they opt out.

Camera-ready Requirements

The final version of all accepted papers will be given one additional page of content (up to 5 pages) so that reviewers’ comments can be taken into account.

Please email any inquiries to mathai.neurips2026@gmail.com.