mlsanityData to 14 Oct 2026

AIMO-2 Winning Solution: Building State-of-the-Art Mathematical Reasoning Models with OpenMathReasoning dataset

arXiv 2504.16891 ↗First posted 23 Apr 2025cs.AI cs.CL cs.LGcited by 174

On the map: LLM reasoning → Mathematical reasoning

Abstract (arXiv)

This paper presents our winning submission to the AI Mathematical Olympiad - Progress Prize 2 (AIMO-2) competition. Our recipe for building state-of-the-art mathematical reasoning models relies on three key pillars. First, we create a large-scale dataset comprising 540K unique high-quality math problems, including olympiad-level problems, and their 3.2M long-reasoning solutions. Second, we develop a novel method to integrate code execution with long reasoning models through iterative training, generation, and quality filtering, resulting in 1.7M high-quality Tool-Integrated Reasoning solutions. Third, we create a pipeline to train models to select the most promising solution from many candidates. We show that such generative solution selection (GenSelect) can significantly improve upon majority voting baseline. Combining these ideas, we train a series of models that achieve state-of-the-art results on mathematical reasoning benchmarks. To facilitate further research, we release our code, models, and the complete OpenMathReasoning dataset under a commercially permissive license.

Key facts. Cited by 174 papers in the map (data through 14 Oct 2026). Placed in Mathematical reasoning, one of 17 directions in LLM reasoning. Builds on 26 of its 39 references that are in the map. First posted on arXiv 23 Apr 2025.

Builds on

26 of 39 references.

Closest ancestors · 7

Same direction, newest first

  1. 2025LIMO: Less is More for Reasoning5 Feb
  2. 2025Leveraging Online Olympiad-Level Math Problems for LLMs Training and Contamination-Resistant Evaluation24 Jan
  3. 2024OpenMathInstruct-2: Accelerating AI for Math with Massive Open-Source Instruction Data2 Oct
  4. 2024Building Math Agents with Multi-Turn Iterative Preference Learning4 Sep
  5. 2024InfinityMATH: A Scalable Instruction Tuning Dataset in Programmatic Mathematical Reasoning9 Aug
  6. 2024OpenMathInstruct-1: A 1.8 Million Math Instruction Tuning Dataset15 Feb
  7. 2024MARIO: MAth Reasoning with code Interpreter Output -- A Reproducible Pipeline16 Jan

Other references · 19

  1. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,459
  2. 2021Training Verifiers to Solve Math Word Problemscited by 10,109
  3. 2025DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learningcited by 6,330
  4. 2022Self-Consistency Improves Chain of Thought Reasoning in Language Modelscited by 5,708
  5. 2021Measuring Mathematical Problem Solving With the MATH Datasetcited by 5,371
  6. 2021RoFormer: Enhanced Transformer with Rotary Position Embeddingcited by 4,703
  7. 2025s1: Simple test-time scalingcited by 1,341
  8. 2022Program of Thoughts Prompting: Disentangling Computation from Reasoning for Numerical Reasoning Taskscited by 1,076
  9. 2024Generative Verifiers: Reward Modeling as Next-Token Predictioncited by 480
  10. 2024Meta-Rewarding Language Models: Self-Improving Alignment with LLM-as-a-Meta-Judgecited by 204
  11. 2023Rethinking Benchmark and Contamination for Language Models with Rephrased Samplescited by 195
  12. 2025ToRL: Scaling Tool-Integrated RLcited by 169
7 more
  1. 2024Generative Reward Modelscited by 148
  2. 2024Critique-out-Loud Reward Modelscited by 106
  3. 2025Agentic Reasoning: A Streamlined Framework for Enhancing LLM Reasoning with Agentic Toolscited by 106
  4. 20251.4 Million Open-Source Distilled Reasoning Dataset to Empower Large Language Model Trainingcited by 49
  5. 2024NeMo-Aligner: Scalable Toolkit for Efficient Model Alignmentcited by 41
  6. 2024Efficient Tool Use with Chain-of-Abstraction Reasoningcited by 33
  7. 2024Direct Judgement Preference Optimizationcited by 33

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 72; 2026: 102.

174 papers.

Sort:

By area:

  1. 2026GLM-5: from Vibe Coding to Agentic Engineeringcited by 587
  2. 2025Kimi K2: Open Agentic Intelligencecited by 567
  3. 2025OpenThoughts: Data Recipes for Reasoning Modelscited by 304
  4. 2025A Survey of Reinforcement Learning for Large Reasoning Modelscited by 180
  5. 2025Llama-Nemotron: Efficient Reasoning Modelscited by 158
  6. 2025Goedel-Prover-V2: Scaling Formal Theorem Proving with Scaffolded Data Synthesis and Self-Correctioncited by 154
  7. 2025REASONING GYM: Reasoning Environments for Reinforcement Learning with Verifiable Rewardscited by 130
  8. 2025NVIDIA Nemotron Nano 2: An Accurate and Efficient Hybrid Mamba-Transformer Reasoning Modelcited by 108
  9. 2025Nemotron 3 Nano: Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Model for Agentic Reasoningcited by 104
  10. 2025VerlTool: Towards Holistic Agentic Reinforcement Learning with Tool Usecited by 96
  11. 2025CWM: An Open-Weights LLM for Research on Code Generation with World Modelscited by 93
  12. 2025AceReason-Nemotron: Advancing Math and Code Reasoning through Reinforcement Learningcited by 87
38 more
  1. 2025AceReason-Nemotron 1.1: Advancing Math and Code Reasoning through SFT and RL Synergycited by 77
  2. 2025Expected Attention: KV Cache Compression by Estimating Attention from Future Queries Distributioncited by 74
  3. 2025Think Only When You Need with Large Hybrid-Reasoning Modelscited by 66
  4. 2025e3: Learning to Explore Enables Extrapolation of Test-Time Compute for LLMscited by 59
  5. 2025rStar2-Agent: Agentic Reasoning Technical Reportcited by 54
  6. 2025Advancing Multimodal Reasoning: From Optimized Cold Start to Staged Reinforcement Learningcited by 53
  7. 2025Nemotron-Cascade: Scaling Cascaded Reinforcement Learning for General-Purpose Reasoning Modelscited by 48
  8. 2025QuestA: Expanding Reasoning Capacity in LLMs via Question Augmentationcited by 44
  9. 2026On Data Engineering for Scaling LLM Terminal Capabilitiescited by 43
  10. 2026OpenResearcher: A Fully Open Pipeline for Long-Horizon Deep Research Trajectory Synthesiscited by 37
  11. 2025CoT-Self-Instruct: Building high-quality synthetic prompts for reasoning and non-reasoning taskscited by 32
  12. 2025InfLLM-V2: Dense-Sparse Switchable Attention for Seamless Short-to-Long Adaptationcited by 32
  13. 2025The Open Proof Corpus: A Large-Scale Study of LLM-Generated Mathematical Proofscited by 30
  14. 2025Front-Loading Reasoning: The Synergy between Pretraining and Post-Training Datacited by 27
  15. 2025Nemotron-Math: Efficient Long-Context Distillation of Mathematical Reasoning from Multi-Mode Supervisioncited by 26
  16. 2025OpenCodeReasoning-II: A Simple Test Time Scaling Approach via Self-Critiquecited by 25
  17. 2025Set Block Decoding is a Language Model Inference Acceleratorcited by 23
  18. 2025Not All Tokens Are What You Need In Thinkingcited by 20
  19. 2025Hybrid Reinforcement: When Reward Is Sparse, It's Better to Be Densecited by 20
  20. 2025Towards Concise and Adaptive Thinking in Large Reasoning Models: A Surveycited by 18
  21. 2025ARES: Multimodal Adaptive Reasoning via Difficulty-Aware Token-Level Entropy Shapingcited by 17
  22. 2025AgentMath: Empowering Mathematical Reasoning for Large Language Models via Tool-Augmented Agentcited by 17
  23. 2026Distribution-Aligned Sequence Distillation for Superior Long-CoT Reasoningcited by 17
  24. 2025QFFT, Question-Free Fine-Tuning for Adaptive Reasoningcited by 15
  25. 2025Don't Pass@k: A Bayesian Framework for Large Language Model Evaluationcited by 15
  26. 2025From Code Foundation Models to Agents and Applications: A Comprehensive Survey and Practical Guide to Code Intelligencecited by 15
  27. 2026On-Policy Delta Distillationcited by 15
  28. 2025Xolver: Multi-Agent Reasoning with Holistic Experience Learning Just Like an Olympiad Teamcited by 13
  29. 2025The Markovian Thinker: Architecture-Agnostic Linear Scaling of Reasoningcited by 13
  30. 2025EntroPIC: Towards Stable Long-Term Training of LLMs via Entropy Stabilization with Proportional-Integral Controlcited by 13
  31. 2025Beyond Exponential Decay: Rethinking Error Accumulation in Large Language Modelscited by 12
  32. 2025The Synergy Dilemma of Long-CoT SFT and RL: Investigating Post-Training Techniques for Reasoning VLMscited by 12
  33. 2025Learning to Refine: Self-Refinement of Parallel Reasoning in LLMscited by 12
  34. 2025Evaluating Parameter Efficient Methods for RLVRcited by 12
  35. 2026AfriqueLLM: How Data Mixing and Model Architecture Impact Continued Pre-training for African Languagescited by 12
  36. 2026Depth-Recurrent Attention Mixtures: Giving Latent Reasoning the Attention it Deservescited by 12
  37. 2026Agentic Proposing: Enhancing Large Language Model Reasoning via Compositional Skill Synthesiscited by 12
  38. 2026The Last AI Built by Humans: Toward Genuine Recursive Self-Improvementcited by 11

50 most cited of 174 All 174 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2025AMO-Bench: Large Language Models Still Struggle in High School Math Competitions
  2. 2025RIMO: An Easy-to-Evaluate, Hard-to-Solve Olympiad Benchmark for Advanced Mathematical Reasoning
  3. 2024Omni-MATH: A Universal Olympiad Level Mathematic Benchmark For Large Language Models
  4. 2025PromptCoT: Synthesizing Olympiad-level Problems for Mathematical Reasoning in Large Language Models
  5. 2025Challenging the Boundaries of Reasoning: An Olympiad-Level Math Benchmark for Large Language Models
  6. 2026AIMO Interpretability Challenge
  7. 2026Escaping the Cognitive Well: Efficient Competition Math with Off-the-Shelf Models
  8. 2025UGMathBench: A Diverse and Dynamic Benchmark for Undergraduate-Level Mathematical Reasoning with Large Language Models