Mathematical Reasoning papers, 2023–2026

Key findings

• Mathematical Reasoning went from 57 papers in 2023 to 577 in 2026, roughly 10x in three years.
• NeurIPS publishes the most Mathematical Reasoning work: 469 papers from 2023 to 2026, 43% of the four-venue total.
• The fastest-growing overlap is with Reinforcement Learning: papers tagged with both rose from 3 in 2023 to 140 in 2026.
• Among the 5 topics under Neurosymbolic AI, it ranks #1 by 2026 paper count; the fastest grower in the group, Differentiable Reasoning, is up 2.5x versus 10x here.

Representative papers

Spread across years and venues so every edition shows up

• NeurIPS 2026A mathematical theory of balancing relational generalization and memorization
• ICML 2026ASyMOB: Algebraic Symbolic Mathematical Operations Benchmark
• ICLR 2026AutoGPS: Automated Geometry Problem Solving via Multimodal Formalization and Deductive Reasoning
• CVPR 2026Affine Perspective-Three-Point Problem
• NeurIPS 2025Are Language Models Efficient Reasoners? A Perspective from Logic Programming
• ICML 2025Autoformulation of Mathematical Optimization Models Using LLMs
• ICLR 2025Advancing Mathematical Reasoning in Language Models: The Impact of Problem-Solving Data, Data Synthesis Methods, and Training Stages
• CVPR 2025Bridging Viewpoint Gaps: Geometric Reasoning Boosts Semantic Correspondence
• NeurIPS 2024A Careful Examination of Large Language Model Performance on Grade School Arithmetic
• ICML 2024Accelerating Legacy Numerical Solvers by Non-intrusive Gradient-based Meta-solving
• ICLR 2024Don't Trust: Verify -- Grounding LLM Quantitative Reasoning with Autoformalization
• CVPR 2024E-GPS: Explainable Geometry Problem Solving via Top-Down Solver and Bottom-Up Generator
• NeurIPS 2023Evaluating and Improving Tool-Augmented Computation-Intensive Math Reasoning
• ICML 2023Automated Search for Conjectures on Mathematical Constants using Analysis of Integer Sequences
• ICLR 2023Differentiable Mathematical Programming for Object-Centric Representation Learning
• CVPR 2023Abstract Visual Reasoning: An Algebraic Approach for Solving Raven's Progressive Matrices
• NeurIPS 2026A Mechanistic Analysis of Looped Reasoning Language Models
• ICML 2026Beyond Theorem Proving: Formulation, Framework and Benchmark for Formal Problem-Solving
• ICLR 2026CogFlow: Bridging Perception and Reasoning through Knowledge Internalization for Visual Mathematical Problem Solving
• CVPR 2026Hilbert-Geo Solving Solid Geometric Problems by Neural-Symbolic Reasoning

All Mathematical Reasoning papers

190 papers, page 2 of 4

51. NeurIPS 2025Demystifying Reasoning Dynamics with Mutual Information: Thinking Tokens are Information Peaks in LLM Reasoning
52. NeurIPS 2025DisCO: Reinforcing Large Reasoning Models with Discriminative Constrained Optimization
53. NeurIPS 2025Diversity-Aware Policy Optimization for Large Language Model Reasoning
54. NeurIPS 2025DLoFT: Gradient-Decoupled Fine-Tuning for Generalizable Long Chain-of-Thought Reasoning
55. NeurIPS 2025Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model?
56. NeurIPS 2025Don’t Think Longer, Think Wisely: Optimizing Thinking Dynamics for Large Reasoning Models
57. NeurIPS 2025DreamPRM: Domain-reweighted Process Reward Model for Multimodal Reasoning
58. NeurIPS 2025DSRF: A Dynamic and Scalable Reasoning Framework for Solving RPMs
59. NeurIPS 2025DuetGraph: Coarse-to-Fine Knowledge Graph Reasoning with Dual-Pathway Global-Local Fusion
60. NeurIPS 2025DyFlow: Dynamic Workflow Framework for Agentic Reasoning
61. NeurIPS 2025DynaAct: Large Language Model Reasoning with Dynamic Action Spaces
62. NeurIPS 2025Efficiently Scaling LLM Reasoning Programs with Certaindex
63. NeurIPS 2025Enigmata: Scaling Logical Reasoning in Large Language Models with Synthetic Verifiable Puzzles
64. NeurIPS 2025EvaLearn: Quantifying the Learning Capability and Efficiency of LLMs via Sequential Problem Solving
65. NeurIPS 2025Evaluating the Inductive Abilities of Large Language Models: Why Chain-of-Thought Reasoning Sometimes Hurts More Than Helps
66. NeurIPS 2025Evolutionary Reasoning Does Not Arise in Standard Usage of Protein Language Models
67. NeurIPS 2025Execution Guided Line-by-Line Code Generation
68. NeurIPS 2025ExPO: Unlocking Hard Reasoning with Self-Explanation-Guided Reinforcement Learning
69. NeurIPS 2025Far from the Shallow: Brain-Predictive Reasoning Embedding through Residual Disentanglement
70. NeurIPS 2025First SFT, Second RL, Third UPT: Continual Improving Multi-Modal LLM Reasoning via Unsupervised Post-Training
71. NeurIPS 2025Flexible Realignment of Language Models
72. NeurIPS 2025From Euler to AI: Unifying Formulas for Mathematical Constants
73. NeurIPS 2025Generalizable Reasoning through Compositional Energy Minimization
74. NeurIPS 2025GPO: Learning from Critical Steps to Improve LLM Reasoning
75. NeurIPS 2025GRIP: A Graph-Based Reasoning Instruction Producer
76. NeurIPS 2025HARDMath2: A Benchmark for Applied Mathematics Built by Students as Part of a Graduate Class
77. NeurIPS 2025HELM: Hyperbolic Large Language Models via Mixture-of-Curvature Experts
78. NeurIPS 2025Heterogeneous Adversarial Play in Interactive Environments
79. NeurIPS 2025How Far Are We from Optimal Reasoning Efficiency?
80. NeurIPS 2025Hybrid Latent Reasoning via Reinforcement Learning
81. NeurIPS 2025Hybrid-Balance GFlowNet for Solving Vehicle Routing Problems
82. NeurIPS 2025Improve Temporal Reasoning in Multimodal Large Language Models via Video Contrastive Decoding
83. NeurIPS 2025Incentivizing Dual Process Thinking for Efficient Large Language Model Reasoning
84. NeurIPS 2025Incentivizing LLMs to Self-Verify Their Answers
85. NeurIPS 2025Incentivizing Reasoning for Advanced Instruction-Following of Large Language Models
86. NeurIPS 2025Ineq-Comp: Benchmarking Human-Intuitive Compositional Reasoning in Automated Theorem Proving of Inequalities
87. NeurIPS 2025IneqSearch: Hybrid Reasoning for Olympiad Inequality Proofs
88. NeurIPS 2025Is PRM Necessary? Problem-Solving RL Implicitly Induces PRM Capability in LLMs
89. NeurIPS 2025Knot So Simple: A Minimalistic Environment for Spatial Reasoning
90. NeurIPS 2025KTAE: A Model-Free Algorithm to Key-Tokens Advantage Estimation in Mathematical Reasoning
91. NeurIPS 2025Large Language Models Think Too Fast To Explore Effectively
92. NeurIPS 2025Latent Refinement via Flow Matching for Training-free Linear Inverse Problem Solving
93. NeurIPS 2025Learning Interestingness in Automated Mathematical Theory Formation
94. NeurIPS 2025Learning to Solve Complex Problems via Dataset Decomposition
95. NeurIPS 2025Learning When to Think: Shaping Adaptive Reasoning in R1-Style Models via Multi-Stage RL
96. NeurIPS 2025Let LRMs Break Free from Overthinking via Self-Braking Tuning
97. NeurIPS 2025LIMOPro: Reasoning Refinement for Efficient and Effective Test-time Scaling
98. NeurIPS 2025Look-Ahead Reasoning on Learning Platforms
99. NeurIPS 2025Lost in Transmission: When and Why LLMs Fail to Reason Globally
100. NeurIPS 2025MathArena: Evaluating LLMs on Uncontaminated Math Competitions

FAQ

How many Mathematical Reasoning papers were published at NeurIPS, ICLR, ICML and CVPR from 2023 to 2026?

1,085 in total: 57 in 2023, 137 in 2024, 314 in 2025, 577 in 2026.

How many Mathematical Reasoning papers are at NeurIPS 2026?

NeurIPS 2026 has 208 Mathematical Reasoning papers, versus 190 the year before.

How many Mathematical Reasoning papers are at ICLR 2026?

ICLR 2026 has 175 Mathematical Reasoning papers, versus 61 the year before.

How many Mathematical Reasoning papers are at ICML 2026?

ICML 2026 has 113 Mathematical Reasoning papers, versus 45 the year before.

How many Mathematical Reasoning papers are at CVPR 2026?

CVPR 2026 has 81 Mathematical Reasoning papers, versus 18 the year before.

Which conference publishes the most Mathematical Reasoning papers?

NeurIPS, with 469 papers from 2023 to 2026.