mlsanityData to 14 Oct 2026

Between Underthinking and Overthinking: An Empirical Study of Reasoning Length and correctness in LLMs

arXiv 2505.00127 ↗First posted 30 Apr 2025cs.CL cs.AIcited by 117

On the map: LLM reasoning → Test-time compute and efficient thinking

Abstract (arXiv)

Large language models (LLMs) are increasingly optimized for long reasoning, under the assumption that more reasoning leads to better performance. However, emerging evidence suggests that longer responses can sometimes degrade accuracy rather than improve it. In this paper, we conduct a systematic empirical study of the relationship between reasoning length and answer correctness. We find that LLMs tend to overthink simple problems, generating unnecessarily long outputs, and underthink harder ones, failing to extend their reasoning when it is most needed. This indicates that models might misjudge problem difficulty and fail to calibrate their response length appropriately. Furthermore, we investigate the effects of length reduction with a preference optimization algorithm when simply preferring the shorter responses regardless of answer correctness. Experiments show that the generation length can be significantly reduced while maintaining acceptable accuracy. Our findings highlight generation length as a meaningful signal for reasoning behavior and motivate further exploration into LLMs' self-awareness in reasoning length adaptation.

Key facts. Cited by 117 papers in the map (data through 14 Oct 2026). Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning. Builds on 30 of its 35 references that are in the map. First posted on arXiv 30 Apr 2025.

Builds on

30 of 35 references.

Closest ancestors · 17

Same direction, newest first

  1. 2025Reasoning Models Can Be Effective Without Thinking14 Apr
  2. 2025L1: Controlling How Long A Reasoning Model Thinks With Reinforcement Learning6 Mar
  3. 2025DAST: Difficulty-Adaptive Slow-Thinking for Large Reasoning Models6 Mar
  4. 2025Self-Training Elicits Concise Reasoning in Large Language Models27 Feb
  5. 2025Towards Thinking-Optimal Scaling of Test-Time Compute for LLM Reasoning25 Feb
  6. 2025The relationship between reasoning and performance in large language models--o3 (mini) thinks harder, not longer21 Feb
  7. 2025CoT-Valve: Length-Compressible Chain-of-Thought Tuning13 Feb
  8. 2025Demystifying Long Chain-of-Thought Reasoning in LLMs5 Feb
  9. 2025s1: Simple test-time scaling31 Jan
  10. 2025O1-Pruner: Length-Harmonizing Fine-Tuning for O1-Like Reasoning Pruning22 Jan
  11. 2024Do NOT Think That Much for 2+3=? On the Overthinking of o1-Like LLMs30 Dec
  12. 2024Token-Budget-Aware LLM Reasoning24 Dec
5 more
  1. 2024Learning How Hard to Think: Input-Adaptive Allocation of LM Computation7 Oct
  2. 2024Adaptive Inference-Time Compute: LLMs Can Predict if They Can Do Better, Even Mid-Generation3 Oct
  3. 2024Make Every Penny Count: Difficulty-Adaptive Self-Consistency for Cost-Efficient Reasoning24 Aug
  4. 2024DynaThink: Fast or Slow? A Dynamic Decision-Making Framework for Large Language Models1 Jul
  5. 2024Escape Sky-high Cost: Early-stopping Self-Consistency for Multi-step Reasoning19 Jan

Other references · 13

  1. 2021Training Verifiers to Solve Math Word Problemscited by 10,109
  2. 2025DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learningcited by 6,330
  3. 2022Self-Consistency Improves Chain of Thought Reasoning in Language Modelscited by 5,708
  4. 2021Measuring Mathematical Problem Solving With the MATH Datasetcited by 5,371
  5. 2025Kimi k1.5: Scaling Reinforcement Learning with LLMscited by 1,210
  6. 2024SimPO: Simple Preference Optimization with a Reference-Free Rewardcited by 963
  7. 2025Towards Reasoning Era: A Survey of Long Chain-of-Thought for Reasoning Large Language Modelscited by 384
  8. 2025From System 1 to System 2: A Survey of Reasoning Large Language Modelscited by 320
  9. 2025Logic-RL: Unleashing LLM Reasoning with Rule-Based Reinforcement Learningcited by 262
  10. 2025When More is Less: Understanding Chain-of-Thought Length in LLMscited by 183
  11. 2024The Impact of Reasoning Step Length on Large Language Modelscited by 153
  12. 2025Optimizing Test-Time Compute via Meta Reinforcement Fine-Tuningcited by 123
1 more
  1. 2025GRPO-LEAD: A Difficulty-Aware Reinforcement Learning Approach for Concise Mathematical Reasoning in Language Modelscited by 82

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 50; 2026: 67.

117 papers.

Sort:

By area:

  1. 2025A Survey of Reinforcement Learning for Large Reasoning Modelscited by 180
  2. 2025A Survey of Context Engineering for Large Language Modelscited by 126
  3. 2025Inverse Scaling in Test-Time Computecited by 50
  4. 2025EDGE-GRPO: Entropy-Driven GRPO with Guided Error Correction for Advantage Diversitycited by 41
  5. 2025OptimalThinkingBench: Evaluating Over and Underthinking in LLMscited by 35
  6. 2025Think or Not? Exploring Thinking Efficiency in Large Reasoning Models via an Information-Theoretic Lenscited by 34
  7. 2025AlphaOne: Reasoning Models Thinking Slow and Fast at Test Timecited by 29
  8. 2025A Comprehensive Survey on Trustworthiness in Reasoning with Large Language Modelscited by 25
  9. 2025Overclocking LLM Reasoning: Monitoring and Controlling Thinking Path Lengths in LLMscited by 22
  10. 2026Think Deep, Not Just Long: Measuring LLM Reasoning Effort via Deep-Thinking Tokenscited by 22
  11. 2025AdaTooler-V: Adaptive Tool-Use for Images and Videoscited by 21
  12. 2025Thinking Fast and Right: Balancing Accuracy and Reasoning Length with Adaptive Rewardscited by 20
38 more
  1. 2025CP-Router: An Uncertainty-Aware Router Between LLM and LRMcited by 20
  2. 2025DiFFPO: Training Diffusion LLMs to Reason Fast and Furious via Reinforcement Learningcited by 19
  3. 2025Towards Concise and Adaptive Thinking in Large Reasoning Models: A Surveycited by 18
  4. 2025The Evolution of Thought: Tracking LLM Overthinking via Reasoning Dynamics Analysiscited by 18
  5. 2025WebCoT: Enhancing Web Agent Reasoning by Reconstructing Chain-of-Thought in Reflection, Branching, and Rollbackcited by 17
  6. 2025Do LLMs Really Forget? Evaluating Unlearning with Knowledge Correlation and Confidence Awarenesscited by 14
  7. 2026InT: Self-Proposed Interventions Enable Credit Assignment in LLM Reasoningcited by 14
  8. 2025General Agentic Planning Through Simulative Reasoning with World Modelscited by 13
  9. 2025The First Impression Problem: Internal Bias Triggers Overthinking in Reasoning Modelscited by 10
  10. 2026DatBench: Discriminative, Faithful, and Efficient VLM Evaluationscited by 10
  11. 2026The Hot Mess of AI: How Does Misalignment Scale With Model Intelligence and Task Complexity?cited by 10
  12. 2026Quantized Reasoning Models Think They Need to Think Longer, but They Do Notcited by 10
  13. 2025Lexical Hints of Accuracy in LLM Reasoning Chainscited by 9
  14. 2025The Shape of Reasoning: Topological Analysis of Reasoning Traces in Large Language Modelscited by 9
  15. 2026Think Shallow, Solve Deep: Controlling Recurrent Dynamics for Reliable Test-Time Depthcited by 9
  16. 2025Which LLMs Get the Joke? Probing Non-STEM Reasoning Abilities with HumorBenchcited by 8
  17. 2025R-Horizon: How Far Can Your Large Reasoning Model Really Go in Breadth and Depth?cited by 8
  18. 2025Towards Understanding Self-play for LLM Reasoningcited by 8
  19. 2025SynapseRoute: An Auto-Route Switching Framework on Dual-State Large Language Modelcited by 7
  20. 2025Self-Aligned Reward: Towards Effective and Efficient Reasonerscited by 7
  21. 2026Probing the Trajectories of Reasoning Traces in Large Language Modelscited by 7
  22. 2026Stop When Reasoning Converges: Semantic-Preserving Early Exit for Reasoning Modelscited by 7
  23. 2025Thinking Short and Right Over Thinking Long: Serving LLM Reasoning Efficiently and Accuratelycited by 6
  24. 2025From Reasoning to Answer: Empirical, Attention-Based and Mechanistic Insights into Distilled DeepSeek R1 Modelscited by 6
  25. 2025RADAR: Reasoning-Ability and Difficulty-Aware Routing for Reasoning LLMscited by 6
  26. 2025LYNX: Learning Dynamic Exits for Confidence-Controlled Reasoningcited by 6
  27. 2026LELA: an LLM-based Entity Linking Approach with Zero-Shot Domain Adaptationcited by 6
  28. 2026Towards a Mechanistic Understanding of Large Reasoning Models: A Survey of Training, Inference, and Failurescited by 6
  29. 2026Towards Efficient Large Language Reasoning Models via Extreme-Ratio Chain-of-Thought Compressioncited by 6
  30. 2026Statistical Early Stopping for Reasoning Modelscited by 6
  31. 2026Mitigating Overthinking in Large Reasoning Language Models via Reasoning Path Deviation Monitoringcited by 6
  32. 2026When Self-Belief Misleads: Active Label Acquisition for Reinforcement Learning with Verifiable Rewardscited by 6
  33. 2025CyclicReflex: Improving Reasoning Models via Cyclical Reflection Token Schedulingcited by 5
  34. 2026Overthinking Loops in Agents: A Structural Risk via MCP Toolscited by 5
  35. 2026LEDGERMIND: Provenance-Constrained Multimodal Agentic Reasoning with a Structured Evidence Ledgercited by 5
  36. 2025Understanding the Role of Training Data in Test-Time Scalingcited by 4
  37. 2025Beyond Token Length: Step Pruner for Efficient and Accurate Reasoning in Large Language Modelscited by 4
  38. 2025CoS: Towards Optimal Event Scheduling via Chain-of-Schedulingcited by 4

50 most cited of 117 All 117 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2025Revisiting Overthinking in Long Chain-of-Thought from the Perspective of Self-Doubt
  2. 2025No Need for Explanations: LLMs can implicitly learn from mistakes in-context
  3. 2025Do LLMs Really Need 10+ Thoughts for "Find the Time 1000 Days Later"? Towards Structural Understanding of LLM Overthinking
  4. 2026Not Too Short, Not Too Long: How LLM Response Length Shapes People's Critical Thinking in Error Detection
  5. 2026When More Thinking Hurts: Overthinking in LLM Test-Time Compute Scaling
  6. 2025Answer-Centric or Reasoning-Driven? Uncovering the Latent Memory Anchor in LLMs
  7. 2025Double-Checker: Enhancing Reasoning of Slow-Thinking LLMs via Self-Critical Fine-Tuning
  8. 2025Measuring Reasoning Utility in LLMs via Conditional Entropy Reduction