mlsanityData to 7 Oct 2026

ProcessBench: Identifying Process Errors in Mathematical Reasoning

arXiv 2412.06559 ↗First posted 9 Dec 2024cs.AI cs.CL cs.LGcited by 277

On the map: LLM reasoning → Reinforcement learning for reasoning

Abstract (arXiv)

As language models regularly make mistakes when solving math problems, automated identification of errors in the reasoning process becomes increasingly significant for their scalable oversight. In this paper, we introduce ProcessBench for measuring the ability to identify erroneous steps in mathematical reasoning. It consists of 3,400 test cases, primarily focused on competition- and Olympiad-level math problems. Each test case contains a step-by-step solution with error location annotated by human experts. Models are required to identify the earliest step that contains an error, or conclude that all steps are correct. We conduct extensive evaluation on ProcessBench, involving two types of models: process reward models (PRMs) and critic models, where for the latter we prompt general language models to critique each solution step by step. We draw two main observations: (1) Existing PRMs typically fail to generalize to more challenging math problems beyond GSM8K and MATH. They underperform both critic models (i.e., prompted general language models) and our own trained PRM that is straightforwardly fine-tuned on the PRM800K dataset. (2) The best open-source model, QwQ-32B-Preview, has demonstrated the critique capability competitive with the proprietary model GPT-4o, despite that it still lags behind the reasoning-specialized o1-mini. We hope ProcessBench can foster future research in reasoning process assessment, paving the way toward scalable oversight of language models.

Key facts. Cited by 277 papers in the map (data through 7 Oct 2026). Placed in Reinforcement learning for reasoning, one of 17 directions in LLM reasoning. Builds on 24 of its 29 references that are in the map. First posted on arXiv 9 Dec 2024.

Builds on

24 of 29 references.

Closest ancestors · 3

Same direction, newest first

  1. 2024Improve Mathematical Reasoning in Language Models by Automated Process Supervision5 Jun
  2. 2023Math-Shepherd: Verify and Reinforce LLMs Step-by-step without Human Annotations14 Dec
  3. 2023Let's Verify Step by Step31 May

Other references · 21

  1. 2024The Llama 3 Herd of Modelscited by 14,192
  2. 2021Training Verifiers to Solve Math Word Problemscited by 9,743
  3. 2023Efficient Memory Management for Large Language Model Serving with PagedAttentioncited by 5,767
  4. 2021Measuring Mathematical Problem Solving With the MATH Datasetcited by 5,203
  5. 2024GPT-4o System Cardcited by 4,467
  6. 2024Qwen2 Technical Reportcited by 2,076
  7. 2016Concrete Problems in AI Safetycited by 1,582
  8. 2024Qwen2.5-Coder Technical Reportcited by 1,151
  9. 2024OlympiadBench: A Challenging Benchmark for Promoting AGI with Olympiad-Level Bilingual Multimodal Scientific Problemscited by 1,099
  10. 2024Qwen2.5-Math Technical Report: Toward Mathematical Expert Model via Self-Improvementcited by 1,057
  11. 2023Large Language Models Cannot Self-Correct Reasoning Yetcited by 965
  12. 2022Self-critiquing models for assisting human evaluatorscited by 372
9 more
  1. 2024Omni-MATH: A Universal Olympiad Level Mathematic Benchmark For Large Language Modelscited by 277
  2. 2024When Can LLMs Actually Correct Their Own Mistakes? A Critical Survey of Self-Correction of LLMscited by 230
  3. 2022Measuring Progress on Scalable Oversight for Large Language Modelscited by 212
  4. 2024LLM Critics Help Catch LLM Bugscited by 138
  5. 2024CriticBench: Benchmarking LLMs for Critique-Correct Reasoningcited by 88
  6. 2024Building Math Agents with Multi-Turn Iterative Preference Learningcited by 66
  7. 2024Is Your Model Really A Good Math Reasoner? Evaluating Mathematical Reasoning with Checklistcited by 48
  8. 2024Towards Scalable Automated Alignment of LLMs: A Surveycited by 33
  9. 2024Yi-Lightning Technical Reportcited by 11

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 2; 2025: 159; 2026: 116.

277 papers.

Sort:

By area:

  1. 2025Search-o1: Agentic Search-Enhanced Large Reasoning Modelscited by 423
  2. 2025The Lessons of Developing Process Reward Models in Mathematical Reasoningcited by 402
  3. 2025Towards Reasoning Era: A Survey of Long Chain-of-Thought for Reasoning Large Language Modelscited by 381
  4. 2025From System 1 to System 2: A Survey of Reasoning Large Language Modelscited by 319
  5. 2025Which Agent Causes Task Failures and When? On Automated Failure Attribution of LLM Multi-Agent Systemscited by 170
  6. 2025PhysBench: Benchmarking and Enhancing Vision-Language Models for Physical World Understandingcited by 147
  7. 2025From LLM Reasoning to Autonomous AI Agents: A Comprehensive Reviewcited by 146
  8. 2025Can 1B LLM Surpass 405B LLM? Rethinking Compute-Optimal Test-Time Scalingcited by 135
  9. 2025VisualPRM: An Effective Process Reward Model for Multimodal Reasoningcited by 134
  10. 2025Process Reward Models That Thinkcited by 118
  11. 2025A Survey of Frontiers in LLM Reasoning: Inference Scaling, Learning to Reason, and Agentic Systemscited by 108
  12. 2025Reward-Guided Speculative Decoding for Efficient LLM Reasoningcited by 106
38 more
  1. 2025PRMBench: A Fine-grained and Challenging Benchmark for Process-Level Reward Modelscited by 98
  2. 2025GenPRM: Scaling Test-Time Compute of Process Reward Models via Generative Reasoningcited by 94
  3. 2024Formal Mathematical Reasoning: A New Frontier in AIcited by 80
  4. 2024AceMath: Advancing Frontier Math Reasoning with Post-Training and Reward Modelingcited by 76
  5. 2025Agent-R: Training Language Model Agents to Reflect via Iterative Self-Trainingcited by 74
  6. 2025InternLM-XComposer2.5-Reward: A Simple Yet Effective Multi-Modal Reward Modelcited by 73
  7. 2025Evaluating Step-by-step Reasoning Traces: A Surveycited by 62
  8. 2025PhysReason: A Comprehensive Benchmark towards Physics-Based Reasoningcited by 58
  9. 2025A Comprehensive Survey of Reward Models: Taxonomy, Applications, Challenges, and Futurecited by 53
  10. 2025Reinforcement Learning from Human Feedbackcited by 53
  11. 2025A Survey on Large Language Models for Mathematical Reasoningcited by 52
  12. 2025Stop Summation: Min-Form Credit Assignment Is All Process Reward Model Needs for Reasoningcited by 49
  13. 2025UGPhysics: A Comprehensive Benchmark for Undergraduate Physics Reasoning with Large Language Modelscited by 45
  14. 2025Evaluating Judges as Evaluators: The JETTS Benchmark of LLM-as-Judges as Test-Time Scaling Evaluatorscited by 42
  15. 2025A Survey of Test-Time Compute: From Intuitive Inference to Deliberate Reasoningcited by 40
  16. 2025SPC: Evolving Self-Play Critic via Adversarial Games for LLM Reasoningcited by 40
  17. 2025VersaPRM: Multi-Domain Process Reward Model via Synthetic Reasoning Datacited by 39
  18. 2025RL Tango: Reinforcing Generator and Verifier Together for Language Reasoningcited by 39
  19. 2025CompassVerifier: A Unified and Robust Verifier for LLMs Evaluation and Outcome Rewardcited by 39
  20. 2025R-PRM: Reasoning-Driven Process Reward Modelingcited by 38
  21. 2025S2R: Teaching LLMs to Self-verify and Self-correct via Reinforcement Learningcited by 37
  22. 2025A Survey of Process Reward Models: From Outcome Signals to Process Supervisions for Large Language Modelscited by 37
  23. 2025Enhancing LLM Reasoning with Iterative DPO: A Comprehensive Empirical Investigationcited by 36
  24. 2025Position: Multimodal Large Language Models Can Significantly Advance Scientific Reasoningcited by 33
  25. 2025Detection and Mitigation of Hallucination in Large Reasoning Models: A Mechanistic Perspectivecited by 32
  26. 2025Beyond Correctness: Harmonizing Process and Outcome Rewards through RL Trainingcited by 29
  27. 2025StepWiser: Stepwise Generative Judges for Wiser Reasoningcited by 28
  28. 2026AgentProcessBench: Diagnosing Step-Level Process Quality in Tool-Using Agentscited by 28
  29. 2025How Is LLM Reasoning Distracted by Irrelevant Context? An Analysis Using a Controlled Benchmarkcited by 26
  30. 2025Where Did It All Go Wrong? A Hierarchical Look into Multi-Agent Error Attributioncited by 25
  31. 2025MathAgent: Leveraging a Mixture-of-Math-Agent Framework for Real-World Multimodal Mathematical Error Detectioncited by 24
  32. 2025Retrieval-Augmented Process Reward Model for Generalizable Mathematical Reasoningcited by 23
  33. 2026Reinforcing Chain-of-Thought Reasoning with Self-Evolving Rubricscited by 23
  34. 2025Step-KTO: Optimizing Mathematical Reasoning through Stepwise Binary Feedbackcited by 22
  35. 2025Premise-Augmented Reasoning Chains Improve Error Identification in Math reasoning with LLMscited by 22
  36. 2025AURORA:Automated Training Framework of Universal Process Reward Models via Ensemble Prompting and Reverse Verificationcited by 22
  37. 2025Uncertainty-Aware Step-wise Verification with Generative Reward Modelscited by 21
  38. 2025VLRMBench: A Comprehensive and Challenging Benchmark for Vision-Language Reward Modelscited by 19

50 most cited of 277 All 277 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2024Physics of Language Models: Part 2.2, How to Learn From Mistakes on Grade-School Math Problems
  2. 2025Right Is Not Enough: The Pitfalls of Outcome Supervision in Training LLMs for Math Reasoning
  3. 2026Unmasking Reasoning Processes: A Process-aware Benchmark for Evaluating Structural Mathematical Reasoning in LLMs
  4. 2025StepMathAgent: A Step-Wise Agent for Evaluating Mathematical Processes through Tree-of-Error
  5. 2025Error Classification of Large Language Models on Math Word Problems: A Dynamically Adaptive Framework
  6. 2026GR-Ben: A General Reasoning Benchmark for Evaluating Process Reward Models
  7. 2025Can Large Reasoning Models Improve Accuracy on Mathematical Tasks Using Flawed Thinking?
  8. 2024Exposing the Achilles' Heel: Evaluating LLMs Ability to Handle Mistakes in Mathematical Reasoning